Copyright is owned by the Author of the thesis. Permission is given for a copy to be downloaded by an individual for the purpose of research and private study only. The thesis may not be reproduced elsewhere without the permission of the Author.
Nutrition and Irrigation Studies with Processing Tomato (Lycopersicon esculentum Mill.)
A thesis presented in partial fulfillrnent of
the requirements for the degree of
Doctor of Philosophy in
Plant Science
at Massey University, Palrnerston North,
New Zealand.
Paul R. Johnstone
2005
Abstract: Improved fertilizer and irrigation management has become increasingly
important for tomatoes (Lycopersicon esculentum Mill.) grown for processing. To
reduce potential nutrient loss to the environment due to excessive supply, fertilizer
recommendations should reflect plant demand determined in an optimal root
environment. An aeroponics experiment examined the effect of low and high nutrient
supply during vegetative growth, fruit development and fruit ripening. The use of
aeroponics in a glasshouse environment allowed control of fertility directly at the root
surface. A further experiment applying aeroponics results was established in the field
using drip-fertigation. Both studies were conducted at Massey University, Palmerston
North. Across experiments, fruit yield was largely determined by vegetative growth in
the 6-8 weeks after transplanting; high fruit yields (> 90 Mg ha-I ) were associated with
improved vegetative growth, and in particular larger leaf area. Mild N deficiency was
the principal cause of poor vegetative growth in low nutrient supply treatments. Higher
yield resulted from greater fruit number. Reinstating adequate fertility after vegetative
growth stopped and fruit number was determined did not increase fruit yield. For
maximum fruit yield, plant uptake of N and K was 9.4 and 1 3 .8 g planr', respectively
(equivalent to approximately 2 1 0 and 3 1 0 kg ha-' at a medium planting density).
Greatest nutrient uptake occurred during fruit development. Where practical, fertilizer
application should be concentrated during fruit growth. Heavy late-season K fertigation
did not increase the soluble solids concentration (SSC) of fruit.
Although offering considerable flexibility in nutrient fertigation, the use of drip
irrigation often results in undesirably low SSC. Late-season irrigation management
strategies to increase fruit SSC without excessive yield loss were subsequently
investigated in drip-irrigated fields. Two experiments were conducted at the University
of California, Davis. Irrigation cutoff prior to fruit ripening reduced fruit set, decreased
fruit size, and increased the incidence of fruit rots, making this approach uneconomical .
Irrigation cutback to 25-50 % of reference evapotranspiration imposed at the onset of
fruit ripening (approximately 6 weeks preharvest) was sufficient to improve fruit SSC
and maintain Brix yields (Mg Brix solids ha-I) compared to the current grower practice
(late cutoff). Irrigation cutbacks imposed during ripening did not cause excessive
canopy dieback, nor were fruit culls or rots increased when the crop was harvested at
commercial maturity. Fruit colour and pH were not adversely affected by irrigation
cutback. Brix monitoring of the earliest ripening fruit (when 30-60 % of the fruit
11
surface shows a colour other than green) can help classify fields as to the severity of
irrigation cutback required to reach desirable fruit SSC at harvest. Combined, these
techniques offer considerable flexibility in managing fields for improved fruit SSC
levels.
111
Acknowledgments:
I would like to express thanks to my doctoral committee, Dr. Mike Nichols, Dr.
Keith Fisher, Dr. David Woolley, Dr. Jeff Reid and Dr. Tim Hartz for their assistance,
interest and patience in helping complete these studies. I am especially grateful to Dr.
Hartz for the opportunities and encouragement he has offered during my studies in
California. My thanks to all of the staff at the Plant Growth Unit, Palmerston North,
and the Vegetable Crops Field Headquarters, Davis, for their assistance with the
management of experimental plots and data collection. I would also like to thank
NORSK-Hydro for financial support of the initial aeroponic study, and Heinz Wattie's
N.Z. Ltd. for supplying tomato seed. I am especially grateful to AGMARDT for
supporting my doctoral tenure. Above all, I thank my friends, family and wife for their
support and encouragement from the early inception of this adventure through to its
more recent completion.
iv
Table of Contents:
Abstract ii
Acknowledgements IV
Table of contents v
List of Tables viii
List of Figures Xl
List of Appendices Xlll
OVERVIEW:
CHAPTER 1: Review of Literature
1.1 1.2 1.3 1.3
Background
Nutrient Management in Processing Tomato Production
Irrigation Management in Processing Tomato Production
Statement of Objectives
CHAPTER 2: 1998-1999 Aeroponic Nutrition Trial
2.1 Materials and Methods
2. 1 . 1 Experimental Site
2. 1 .2 Production Details
2. 1 .3 Experimental Design
2. 1 .4 Data Collection
2. 1 .5 Statistical Analysis
2.2 Results
2.2. 1 Fertility Index
2.2.2 Destructive Whole-Plant Sampling
2.2.3 Nutrient Analysis
2.2.4 Fruit Yield and Quality Measurements
2.2.5 Nutrient Uptake
2.3 Discussion
2.3. 1 Effect of Fertility on Plant Biomass and Growth
2.3.2 Effect of Fertility on Leaf Nutrient Concentrations
V
1
2
2
6
1 1
1 2
1 2
1 5
1 6
2 1
23
24
39
44
5 1
58
60
2.3.3 Effect of Fertility on Fruit Yield and Quality
2.3 .4 Effect of Fertility on Nutrient Uptake
62
67
CHAPTER 3: 1999-2000 Field Nutrition Trial
3.1 Materials and Methods
3 . 1 . 1
3 . 1 .2
3 . 1 .3
3. 1 .4
3. 1 .5
Experimental Site
Production Details
Experimental Design
Data Collection
Statistical Analysis
7 1
72
73
76
77
3.2 Results
3.2. 1 Leaf Nutrient Sampling
3.2 .2 Destructive Whole-Plant Biomass Sampling
3 .2.3 Fruit Yield and Quality Measurements
77
80
82
3.3 Discussion
3.3.1 Effect of Fertilizer Application on Leaf Nutrient Concentrations 89
3 .3 .2 Effect of Fertilizer Application on Whole-Plant Biomass 9 1
3 .3 .3 Effect of Fertilizer Application on Fruit Yield and Quality 92
CHAPTER 4: 2001 and 2002 Irrigation Management Trials
4.1 Materials and Methods
4.1.1 Experimental Site
4. 1 .2 Production Details
4. 1 .3 Experimental Design
4. 1 .4 Data Collection
4. 1 .5 Statistical Analysis
4.2 2001 Results
4.2 . 1 Plant Canopy Cover Measurements
4 .2.2 Plant Water Status
4.2.3 Soil Water Status
4.2.4 Fruit Yield and Quality Measurements
4.3 2002 Results
VI
96
96
98
1 00
1 03
1 04
1 05
1 07
1 09
4.3 . 1 Plant Water Status
4.3.2 Soil Water Status
4.3.3 Fruit Yield and Quality Measurements
4.4 Discussion
4.4. 1 Effect of Water Application on Plant Canopy Cover
4.4.2 Effect of Water Application on Plant Water Status
4.4.3 Effect of Water Application on Soil Water Status
1 1 7
1 20
1 22
1 29
1 30
1 32
4.4.4 Effect of Water Application on Yield and Quality Measurements 1 33
CHAPTER 5: Summary
5.1 1998-1999 Aeroponic Nutrition Trial
5 . 1 . 1 Effect of Fertility on Plant Growth
5. 1 .2 Effect of Fertility on Leaf Nutrient Concentrations
5 . 1 .3 Effect of Fertility on Fruit Yield and Quality
5 . 1 .4 Effect of Fertility on Nutrient Uptake
5.2 1999-2000 Field Nutrition Trial
5.2 . 1 Effect of Fertility on Leaf Nutrient Concentrations
5 .2.2 Effect of Fertility on Plant Growth
5 .2.3 Effect of Fertility on Fruit Yield and Quality
5.3 2001 and 2002 UCD Irrigation Management Trials
5.3 . 1 Effect of Water Application on Plant Canopy Cover
5 .3.2 Effect of Water Application on Plant Water Status
5 .3 .3 Effect of Water Application on Soil Water Status
5 .3 .4 Effect of Water Application on Yield and Quality
CHAPTER 6: Conclusions
6.1 6.1
Plant Nutrition Studies
Irrigation Management Studies
LITERATURE CITED:
APPENDICES:
List of Tables:
vu
1 39
1 40
1 4 1
1 42
1 43
1 44
1 44
146
1 46
1 47
148
1 5 1
1 52
1 54
1 67
Table 2. 1 . Production and phenological dates, 1 998- 1 999. 1 3
Table 2.2. Factorial treatment design under aeroponics between 1 6-92 1 5
DAT.
Table 2.3. Destructive plant sampling between 8-92 DAT relative to 1 8
fertility treatments.
Table 2.4. Standard orthogonal contrasts made during each period of 22
plant growth.
Table 2 .5 . Fertility regime, fertility index (FIN) and adjusted fertility 24
index (F INAdj) between 1 6-92 DA T.
Table 2.6. Effect of fertility regime on total above-ground dry biomass 26
(Tdw) accumulation at the end of each defined growth period.
Table 2.7. Effect of fertility regime on leaf dry biomass (L dw) 28
accumulation at the end of each defined growth period.
Table 2.8. Effect of fertility regime on stem dry biomass (Sdw) 3 1
accumulation at the end of each defined growth period.
Table 2.9. Effect of fertility regime on fruit dry biomass (Fdw) 33
accumulation at the end of each defined growth period.
Table 2. 1 0. Effect of fertility regime on root dry biomass (RIw) 35
accumulation at the end of each defined growth period.
Table 2 . 1 1 . Effect of fertility regime on leaf area (LA) accumulation at 37
the end of each defined growth period.
Table 2. 1 2 . Effect of fertility regime on relative growth rate (RGR) of 38
plants across each defined period of growth.
Table 2. 1 3 . Effect of fertility regime on net assimilation rate (NAR) and 39
leaf area ratio (LAR) of plants across each defined period of growth.
Table 2. 1 4. Effect of fertility regime on seasonal N concentration (%) in 40
YML tissue.
Table 2. 1 5 . Effect of fertility regime on seasonal P concentration (%) in 42
YML tissue.
Table 2. 1 6 . Effect of fertility regime on seasonal K concentration (%) in 43
YML tissue.
Table 2. 1 7. Effect of fertility regime on tomato yield and fruit quality at 44
Vlll
92 DAT.
Table 2. 1 8 . Effect of fertility regime on titratable acidity (TA) and 50
lycopene content of fruit at 92 DA T.
Table 2 . 1 9. Effect of fertility regime on N, P, K, Ca and Mg 5 1
concentration (%) in fruit tissue at 92 DAT.
Table 2.20. Effect of fertility regime on cumulative N uptake (total , leaf, 53
stem and fruit) at the end of each defined period of growth.
Table 2.2 1 . Effect of fertility regime on cumulative P uptake (total , leaf, 55
stem and fruit) at the end of each defined period of growth.
Table 2 .22. Effect of fertility regime on cumulative K uptake (total , leaf, 57
stem and fruit) at the end of each defined period of growth.
Table 3 . 1 . Soil fertility status of the top 30 cm depth. 72
Table 3.2. Production and phenological dates, 1999-2000. 73
Table 3 .3 . Optimal N and K treatments ( 1 .0x) applied during vegetative 75
development, fruit development and fruit ripening.
Table 3 .4. Effect of fertilizer application rate on YML N concentration 78
(%) at early bloom, full bloom and early fruit ripening.
Table 3 .5 . Effect of fertilizer application rate on YML K concentration 80
(%) at early bloom, full bloom and early fruit ripening.
Table 3 .6. Effect of fertilizer application rate on plant dry biomass (total, 8 1
leaf and stem) and leaf area at 55 DAT.
Table 3 .7. Effect of fertilizer application rate on tomato yield and fruit at 83
1 30 DAT.
Table 3 .8 . Effect of fertilizer application rate on the concentration of N, 87
P and K (%) in fruit at 1 30 DAT.
Table 3 .9. Effect of fertilizer application rate on removal of N, P and K 89
in marketable fruit at 1 30 DAT.
Table 4. 1 . Fertilizer application rates of N, P and K, 200 1 and 2002. 97
Table 4.2. Production and phenological dates, 200 1 and 2002. 97
Table 4.3. Description of irrigation treatments, 2001 and 2002. 99
Table 4.4. Irrigation treatment schedule during the final 60 days 1 04
preharvest.
Table 4.5 . Effect of irrigation treatment on plant canopy cover. 1 04
IX
Table 4.6. Effect of irrigation treatment on 'VStem measured at 0700 HR 1 07
and 1 300 HR.
Table 4.7. Effect of irrigation treatment on available soil moisture 1 09
(AS M) and soil matric potential ('VSoil) across the final six weeks
preharvest.
Table 4.8. Effect of irrigation treatment on tomato yield and fruit quality 1 1 0
at harvest.
Table 4.9. Irrigation treatment schedule during the final 50 days 1 1 7
preharvest.
Table 4. 1 0. Effect of irrigation treatment on 'VStem measured at 0700 HR 1 20
and 1 300 HR.
Table 4. 1 1 . Effect of irrigation treatment on available soil moisture 1 22
(ASM) and soil matric potential ('VSoil) across the final four weeks
preharvest.
Table 4. l 2. Effect of irrigation treatment on tomato yield and fruit 1 23
quality at harvest.
x
List of Figures:
Fig. 2 . 1 a. Early-season vegetative growth at 2 1 DAT. 1 4
Fig. 2 . 1 b . Mid-season flowering and fruit development at 4 2 DAT. 1 4
Fig. 2.2. Root growth inside the aeroponic tank at 72 DAT. 1 6
Fig. 2.3 . Effect of fertility regime on seasonal total plant dry biomass 25
(Tdw) accumulation.
Fig. 2.4. Effect of fertility regime on seasonal leaf dry biomass (Ldw) 27
accumulation.
Fig. 2 .5 . Effect of fertility regime on seasonal stem dry biomass (Sdw) 30
accumulation.
Fig. 2.6. Effect of fertility regime on seasonal fruit dry biomass (Fdw) 32
accumulation.
Fig. 2.7. Effect of fertility regime on seasonal root dry biomass �w) 34
accumulation.
Fig. 2.8 . Effect of fertility regime on seasonal leaf area (LA) expansion. 36
Fig. 2 .9. Relationship between fertility index (FIN) and total fruit yield at 45
92 DAT.
Fig. 2 . 1 O. Relationshi'p between fertility index (FIN) and fruit soluble 48
solids concentration (SSC) at 92 DAT.
Fig. 2 . 1 1 . Relationship between fertility index (FIN) and actual Brix yield 49
(a) and adjusted Brix yield (b) at 92 DAT.
Fig. 3 . 1 a. Early-season vegetative growth at 26 DAT. 74
Fig. 3 . 1 b. Mid-season vegetative growth at 55 DAT. 74
Fig. 3 .2. Effect of fertilizer application rate on the total yield of 'M or se' 84
and 'H225 ' at 1 30 DAT.
Fig. 4. 1 . Early-season vegetative growth, 2001 . 98
Fig. 4.2. Effect of sampling technique on plant water potential ("') in 1 05
control plots at 0700 HR.
Fig. 4.3. Effect of sampling technique on plant water potential ("') in 1 06
control plots (a) and severe stress plots (b) at 1 300 HR.
Fig. 4.4. Effect of irrigation treatment on soil volumetric water content 108
(Bv) in the 0-60 cm (a) and 60- 1 20 cm (b) depth ranges.
xi
Fig. 4 .5 . Effect of irrigation treatment on total (a) and marketable (b) 1 1 1
fruit yield.
Fig. 4.6. Effect of irrigation treatment on fruit soluble solids 1 1 2
concentration (SSC).
Fig. 4.7. Effect of irrigation treatment on Brix yield. Treatment means 1 1 3
are shown.
Fig. 4.8. Effect of irrigation treatment on blended fruit colour. 1 14
Fig. 4.9. Effect of irrigation treatment on mean marketable fruit mass 1 1 5
(MFf). Fig. 4. 1 0. Effect of irrigation treatment on yield components at maturity. 1 1 6
Fig. 4. 1 1 . Effect of sampling technique on plant water potential (0/) in 1 1 8
control plots at 0700 HR.
Fig. 4. 1 2 . Effect of sampling technique on plant water potential (0/) in 1 1 9
control plots (a) and severe stress plots (b) at 1300 HR.
Fig. 4. 1 3 . Effect of irrigation treatment on soil volumetric water content 1 2 1
(By) in the 0-60 cm (a) and 60- 1 20 cm (b) depth ranges.
Fig. 4. 1 4. Effect of irrigation treatment on total (a) and marketable (b) 1 24
fruit yield.
Fig. 4. 1 5 . Effect of irrigation treatment on fruit soluble solids 1 25
concentration (SSe). Fig. 4. 1 6. Effect of irrigation treatment on Brix yield. 1 26
Fig. 4. 1 7. Effect of irrigation treatment on blended fruit colour. 127
Fig. 4. 1 8 . Effect of irrigation treatment on mean marketable fruit mass 1 28
(MFf). Fig. 4. 1 9. Effect of irrigation treatment on yield components at maturity. 1 29
xii
List of Appendices:
Appendix 1. Summary of low (0.7 dS m-I ) and high (2.0 dS m-I) fertility 1 68
treatments, 1 998-1 999 aeroponic trial .
Appendix II-a. Minimum and maximum daily air temperature during 1 69
summer trial, 1 999-2000.
Appendix II-b. Mean daily soil temperature at 1 5 cm during summer 1 69
trial, 1 999-2000.
Appendix Il-c. Daily pan evaporation (Ep) during summer trial, 1 999- 1 70
2000.
Appendix Il-d. Daily rainfall during summer trial, 1 999-2000. 1 70
Appendix Ill-a. Daily reference evapotranspiration (ETo) during summer 1 7 1
trials, 200 1 (a) and 2002 (b).
Appendix Ill-b. Minimum and maximum daily air temperature during 1 72
summer trials, 2001 (a) and 2002 (b).
Appendix Ill-c. Mean daily soil temperature at 1 5 cm during summer 1 73
trials, 2001 (a) and 2002 (b).
Appendix IV. Managing fruit soluble solids with late-season deficit 1 74
irrigation in drip-irrigated processmg tomato production
(HortScience 40: 1 857- 1 86 1 ).
Xlll
OVERVIEW:
Environmental stewardship in agriculture has become increasingly important,
particularly as the extent to which poor fertilizer and irrigation management can
contribute to environmental pollution is revealed. The trend towards improved
stewardship has been further fueled by market demand for "eco-friendly" production
practices. Large retailers and processors want to certify products as being produced
using environmentally-sound techniques; in as much, growers are being encouraged to
adopt production practices that can limit damage to the land.
The agriculture sector in general must therefore continually refine crop
management techniques and ensure that appropriate technology is incorporated where
possible. A challenge in this process has been balancing what is agronomically
acceptable and environmentally desirable; practices that are advantageous to the grower
are not always beneficial to the environment, and vice-versa. One technology with the
potential to address both agronomic and environmental concerns is drip irrigation and
fertigation. By applying nutrients and water directly at the root surface, general
management efficiencies can be improved, and the potential for runoff and leaching to
the greater environment minimized. In recent years drip technology has increased in
use in many agricultural sectors, including the processing tomato industry overseas.
Although crop nutrition and water management can be improved with drip technology,
traditional fertilizer and irrigation practices must first be calibrated to suit this approach.
To address this issue, a series of four experiments were conducted to improve the
management of nutrients and water for drip-irrigated processing tomato.
1
CHAPTER 1: Review of Literature
1.1 Background
The global processing tomato industry has traditionally focused on the
production of bulk paste (for use in ketchup, puree, spaghetti sauce, pizza sauces, soups,
juices, etc.), although demand for whole peeled, or peeled and diced fruit is increasing.
Worldwide, in excess of 30,000,000 Megagram's (Mg) of fruit are produced annually.
Of this, over 30 % is grown in California, making it by far the largest supplier of
processed tomato product; other notable suppliers include Italy, China, Spain, Turkey
and Brazil. By comparison, New Zealand (N.Z.) produces < 0.5 % of the global supply,
most of which is consumed domestically. The constant in all markets is that processors
demand high quality product at cheap prices; margins on production continue to tighten
in most markets. Additionally, other production constraints have emerged. In particular, environmental stewardship continues to be the focus of renewed interest.
Two areas of particular agronomic interest in improving short-term and long-tenn
sustainability are nutrient and irrigation management. However, the high economic
value of processing tomato often requires high-input production practices to maximize
fruit yield and quality.
1.2 Nutrient Management in Processing Tomato Production
To ensure adequate nutrient availability during growth and development,
fertilizer applications often exceed critical plant requirements; many growers use
generalized fertilizer programs for nitrogen (N), phosphorus (P) and potassium (K),
without tailoring them to field-specific conditions (Hartz, 2002a). While excessive
fertilization generally does not cause serious agronomic problems, it can be a significant
contributor to elevated nutrient concentrations in surface and ground waterways (Hartz,
2002a; Pote et al. , 1 996; Rahn, 2002). Soil test measures of N and P have been
positively correlated with nutrient concentrations in field runoff and leachate from
irrigation or rainfall (Blackmer, 1 987; Hartz and Johnstone, 2005; McDowell and
Sharpley, 200 1 ; Sharpley, 1 995, 1997). The negative effect of high levels of soluble
nutrients on the eutrophication of water bodies has been well documented (Keeney and
2
Follett, 1 99 1 ; Pote et al. , 1 996; Sims, 1 998); nitrate contamination may also present a
health risk in drinking water drawn from sub-surface aquifers.
Despite elevated nutrient availability in many soils, growers continue to apply
fertilizers to maximize tomato growth during establishment; early growth is almost
exclusively vegetative, as the plant canopy and root structure are fonned (Picken et aI. ,
1 986). However, once fruit set begins and assimilates are diverted to developing fruit,
most new vegetative growth stops (Halbrooks and Wilcox, 1 980); this is characteristic
of the determinate growth pattern in processing tomato. Establishing a strong
vegetative framework is therefore essential to maximize the interception of radiation for
photosynthetic activity (Greenwood, 200 1 ; Sinclair and Horie, 1 989; Tei et al. , 2002).
The main source of assimilates for fruit growth comes from the leaf (Ho and Hewitt,
1 986), so limiting assimilate supply can reduce fruit growth and subsequent yield
(Schaffer et al. , 1 999). Stevens ( 1 976) suggested that there is rarely an excess of
assimilate supply in processing tomatoes, because fruit set is compact and leaf areas
invariably smaller than with indeterminate fresh-market tomatoes. A compact fruit set
is necessary to allow the maximum fruit yield to be ripe for a once-over destructive
harvest.
Of the nutrients required for growth, high fruit yield is most commonly limited
by N deficiency; this reflects the relative ease with which N can be leached out of the
active root zone with heavy rainfall or excessive irrigation and the overall importance of
N in vegetative tissue (N is a key constituent of chlorophyll, the photosynthetic unit).
Where N limits plant growth, yield is invariably reduced (Adams, 1 986; Cavero et aI. ,
1 997, 1 998; Cerne, 1990; Clark et al. , 1 999; Colla et aI. , 1 999, 200 1 ; Geisenberg and
Stewart, 1 986); the degree of decline varies considerably depending on the extent and
timing of deficiency. Yield decline has often been associated with low fruit number,
apparently reflecting an N limitation on fruit set. Competition for assimilates can cause
flowers to abort, as supplies are preferentially directed to fruit that have already set (Ho,
1 984; Wardlaw, 1 990). Julian (1 990) also provided evidence of this effect, suggesting
that most fruit set on the first 1 8 trusses of a plant; later trusses contributed very little to
fruit yields. Similar trends were reported by Renquist and Reid ( 1 998). Assimilate
supply can also affect mean fruit mass. Fruit size is affected by the number of cells
originating from division processes in the first week following anthesis; the remainder
of fruit growth consists almost entirely of cell elongation (Gillaspy et al. , 1 993). Baldet
et al. (2002) suggested that sugar deprivation limited cell division, reducing overall fruit
3
size. High fruit set can also decrease mean fruit mass (Colla et al., 1 999), although in
most instances, yields are not reduced; fruit yield is a function of fruit number and
individual mass (Ho and Hewitt, 1 986).
For maximum fruit yield and quality, California growers commonly apply
between 1 40-280 kg N ha-! (Hartz and Miyao, 1 997); despite such norms, individual
grower-operations can vary considerably in nutrient management practices based on a
number of perceived production and environmental constraints (soil type, cropping
history, humid/arid conditions, etc.). Some research has suggested considerably lower
fertilizer applications are satisfactory to maximize fruit yield and quality; Krusekopf et
al. (2002) found that although grower N application varied among 1 0 commercial sites
from 1 40-273 kg N ha-I , no more than 1 00 kg N ha-! was necessary in any field to
maximize tomato yield or quality. In these trials, excess N fertilization by the grower
not removed in the harvested fruit was presumably prone to environmental loss during
the winter, as suggested by Rahn (2002). Others have concluded higher rates are
justified (Ceme, 1 990; Dadomo et al., 1994a; Dumas, 1990; Halbrooks and Wilcox,
1 980; Hills et al., 1983); in some instances supply > 250 kg N ha-! was required for
maximum fruit yield. Presumably many experimentally-determined differences relate
to the growing environment, original N status of the soil, production practices and
cultivar selection.
Both P and K can also limit tomato growth (Besford and Maw, 1975; Biddinger
et al. , 1 998; Fontes and Wilcox, 1 984; Hartz et al. , 1 998, 200 1 , 2005 ; Lingle and
Lorenz, 1 969). However, in many agricultural soils P and K availability is considered
to be satisfactory for high fruit yields (Hartz and Miyao, 1 997; Strand, 1 998) due to
high native fertility, or decades of intensive production where fertilizer application rates
have exceeded crop removal. Hartz and Miyao ( 1 997) suggested that a soil test value
(bicarbonate-extractable) of 1 2- 1 5 mg kg-! P is generally interpreted as sufficient; in
fields with soil test values less than 1 2 mg kg-I P fertilization may increase fruit yield,
while above 1 5 mg kg- I applications are not likely to be beneficial. However, the
current grower practice in California is to apply on average 30-60 kg P ha-I (Hartz and
Miyao, 1 997), largely irrespective of soil test values. In part, this reflects the grower
perception that standard chemical extraction procedures may not accurately measure
plant-available forms of P as suggested by Sims et at. (2000). Also, many growers
apply P fertilizer preplant to guarantee adequate availability during the sensitive period
of establishment; poor availability at this time can cause plants to be stunted, reducing
4
growth potential and therefore yield. This problem is often aggravated by spring
plantings, when low soil temperature can limit the conversion of P to plant-available
fonns (Brady, 1 990); Johnstone et al. (2005a) found that for each 1 0 °C decrease in soil
temperature bioavailable P was reduced by approximately 40 %.
Research using traditional preplant and side-dressing fertilization techniques has
found that soils with an exchangeable K value < 0.35 cmol kg-I may respond to K fertilization (Hartz et al. , 2001 ), although this threshold may be higher when K is
fertigated by buried drip lines (Hartz et al. , 2005). Improved efficiency under drip may
reflect the limited movement of K in many soils (fertigation places K directly in the
active root zone), less fixation due to more uniform moisture content (Cassman et al. ,
1 990) and greater availability from frequently increasing soil solution K concentration
(Hartz et al. , 2005). California growers seldom apply more than 1 1 0 kg K ha-I (Hartz
and Miyao, 1 997), although K uptake associated with high fruit yield can exceed 300 kg
K ha- I (Widders and Lorenz, 1 979); comparatively low K applications reflect the ability
of many soils to supply large amounts of K (due to the previous cropping history or
high natural fertility).
A positive relationship between fruit soluble solids concentration (SSC) and K fertilization has been reported in several studies (Dumas, 1 990; Lachover, 1 972).
However, these results have not been widely corroborated. In a survey of 1 40
commercial processing tomato fields in California, Hartz et al. ( 1 999) concluded that
the impact of soil or plant K status on fruit SSC was minor. Additionally, in field trials
investigating various application amounts (0-800 kg K ha- I ), the timing of application
(flowering, fruit set and early fruit ripening) and methods of applications (fertigation,
foliar sprays, preplant and sidedressed), Hartz et al. (200 1 , 2005) found no relationship
between fruit SSC and soil K status or K fertilization. Greater K nutrition has
ameliorated some fruit colour disorders (Hartz et al. , 1 999, 200 1 , 2005; Picha and Hall ,
1 98 1 ; Trudel and Ozbun, 1 970, 1 97 1 ), although even at very high application rates has
not eliminated them. In many instances, K fertilization to maximize fruit colour
required rates far higher than that required for maximum fruit yield; Hartz et al. (2005)
concluded that to justify such applications a grower would need to recover the cost of
the fertilizer in higher yield and/or in a price premium for improved fruit quality.
Many growers continue to adopt a liberal approach to fertilizer management,
applying too much rather than too little. Although some growers use monitoring tools
(including soil testing and leaf analysis) to help guide fertilizer decisions, many do not;
5
since fertilizer cost is a small component of overall crop production costs, some growers
pay insufficient attention to crop fertility issues. To reduce potential environmental
risk, current practices can be refined. Issues of sustainability have increasingly become
the focus of agricultural research and extension (Hartz, 2002a). However, most
research on processing tomato nutrition has been conducted in the field where control of
fertility is less precise. McDonald et al. ( 1 996) suggested it is desirable to manipulate
nutrient supply directly at the root surface, eliminating any buffering capacity of a soil
volume. Soil type can affect the availability of applied nutrients in such experiments.
Subsequently, applied fertilizer treatments may not accurately represent nutrient
availability to the plant. Current fertilizer norms may therefore be inaccurate.
An alternative to field determination is aeroponics, which offers the potential to
study plant nutrition in an optimal root environment, where neither water nor oxygen
are limited or influenced by intrinsic soil properties (Barak et al., 1 996; Waisel, 1 996).
Eymar et al. (200 1 ) concluded that solution culture can improve the accuracy of
nutritional studies by reducing external factors; outcomes may therefore represent the
optimum conditions in which the crop is grown, becoming a useful standard for
companson. Additionally, conventional fertilization techniques rely on prep 1 ant
applications and one or more side-dressings made mid-season. A more effective means
to match plant uptake with supply is to apply nutrients with irrigation water; fertigation
by subsurface drip irrigation allows small amounts of nutrients to be supplied directly to
the active root zone at will (Hartz and Hochmuth, 1 996). At identical seasonal
applications of N and K, continuous drip fertigation improved fruit yield compared to
traditional preplant and side-dress methods (Giiler et al. , 2002). Closely matching
supply with plant uptake is also integral to improving fertilizer management (Dumas,
1 990; Hartz, 2005 ; Stark et al., 1 983); such practices can decrease the potential for
nutrient loss to the environment by reducing availability during periods of low plant
demand.
1.3 Irrigation Management in Processing Tomato Production
Successful irrigation management is very important in maximizing plant
productivity; where water is limiting, yield and quality of agricultural crops can be
greatly affected (Fereres et aI. , 2003). In most environments, processing tomatoes are
irrigated by furrow methods. Irrigation amounts of 50- 1 00 mm are typically applied
6
every 7-1 4 days during plant growth and fruit development (Hartz and Miyao, 1 997),
with significant soil drying occurring between irrigations (Hanson et al. , 2000a).
Conventional irrigation practices can be subject to poor distribution uniformity (a
measure of how uniformly water is applied across the field); Hanson ( 1 995) reported
that furrow irrigation frequently achieved a uniformity of < 75 %. Poor uniformity may
result from insufficient land levelling, long furrow lengths and slow infiltration rates.
In addition to being wasteful and increasing the risk of nutrient loss to the environment
(due to leaching and/or surface runoff), under- and over-watering can also reduce
tomato yield; Hartz (2002b) suggested that more yield potential is commonly lost due to
improper water management than to any other factor under grower control.
By comparison, drip irrigation offers the potential for improved water
management by eliminating many engineering and cultural constraints that complicate
conventional irrigation methods (Hartz, 1 996). Hartz (2005) noted that well-designed
drip systems can reach a distribution uniformity of > 90 %. Drip irrigation can also
reduce or eliminate surface runoff, deep percolation and soil evaporation; water usage
can be lowered by as much 50 % of that required with conventional irrigation methods
(Phene, 1 999), although in most fields water conservation is typically less, within the
range of 1 0-30 % (depending on the growers ' previous irrigation practices).
Drip irrigation minimizes plant water stress during growth and development by
supplying frequent, small-volume irrigations (Hanson et al. , 2003 ; Hartz, 1 996; Phene,
1999); irrigation amounts of 10-15 mm are commonly applied every 2-3 days with drip.
In well-managed fields, conversion to drip irrigation has typically increased processing
tomato fruit yield by 1 0-25 % or more compared to furrow irrigation (Hartz, 200 1 ).
Subsequently, the use of drip irrigation in processing tomato production has increased
rapidly; in 2005, more than 1 6,000 ha of land was estimated to be drip-irrigated in
California (T.K. Hartz, personal communication), representing approximately 1 5 % of
the total land farmed in this market. Despite this trend, drip irrigation is not currently
used on a large scale by N.Z. growers. Although drip systems can be costly to establish,
increased productivity and savings in water, energy and labour have generally recouped
such expenses within 2-3 years of installation (Linden, 2003); buried drip tapes last for
5-7 years, with the pump equipment and filtration systems 1 0- 1 5 years. As competition
for water among agricultural, industrial and urban uses increases, options that reduce
growers' water dependency appear likely to continue to gain popularity.
7
However, some processors remain hesitant of drip irrigation because fruit SSC
has generally been lower than that achieved with furrow irrigation; drip-irrigated fields
have commonly been 0.2-0.5 °Brix lower than furrow averages (Hartz, 200 1 ).
Acceptable fruit SSC levels are typically > 5.0 °Brix. Because factory processing
involves evaporating of unnecessary water, fruit with low SSC reduce both paste yield
per unit of fresh fruit and overall processing efficiency. Even small improvements in
fruit SSC are important to processors because the savings can be substantial (Renquist
and Reid, 200 1 ). Linden (2004) reported that each 0. 1 °Brix decrease in SSC costs
California processors approximately US$ I .30 per Mg of raw product in additional
processing and handling; at an average commercial yield of 90 Mg ha-I, a decline in
fruit SSC of between 0.2-0.5 °Brix would increase processing costs by approximately
US$230-580 ha- I . Growers who consistently deliver fruit with low SSC may risk
contracts in competitive markets where supply is high. Some processors have
implemented price premiums for fruit with higher SSC, providing an incentive to
manage for improved fruit quality. Management techniques that consistently increase
fruit SSC in drip-irrigated fields with minimal yield loss are therefore required.
Irrigation management affects both tomato yield and SSC (Cahn et al., 200 1 ,
2003 ; Calado et aI., 1 990; Dumas et aI., 1 994; Lowengart-Aycicegi et al. , 1 999; May et
al. , 1 990; May and Gonzales, 1 999; Murray, 1 999; Renquist and Reid, 200 1 ). All
reported that imposing soil moisture stress during fruit sizing and ripening reduced yield
while increasing SSc. Poor fruit SSC in drip-irrigated fields has primarily resulted
from insufficient water stress as fruit ripen. Water stress increases plant water potential
(Mitchell et al., 1 99 1 a; Renquist and Reid, 200 1 ; Rudich et aI. , 1 98 1 ), which reduces
the movement of solutes into developing fruit; this decreases fruit expansion and results
in higher sugar concentrations (Ehret and Ho, 1 986). Phene et al. ( 1 986) found that in
processing tomato, plant water use during fruit ripening is only 80-90 % of reference
evapotranspiration (ETo) due to natural leaf senescence and declining water loss; ETo
refers to water loss from a standardized crop, commonly closely-clipped, actively
growing grass. This range reflected plants with strong vigour; actual water use may be
even lower in fields with poor vigour or incomplete canopy cover. Irrigation
management for improved fruit SSC therefore needs to account for this decline in water
use, and also include an additional degree of water deficit to ensure plants are
sufficiently stressed late in the season.
8
A common management approach to increasing fruit SSC in drip-irrigated fields
has been full irrigation cutoff implemented during fruit ripening. With the determinate
cultivars currently used, fruit ripening typically encompasses the final 6 weeks
preharvest. During this period, fruit ripen at an average rate of 2-3 % of fruit reaching
the pink stage of maturity each day. Although irrigation cutoff 6 weeks or more
preharvest increases fruit SSC, it can also dramatically reduce fruit set (Atherton and
Othman, 1 983; Colla et al. , 1 999; May et al. , 1 990; Wudiri and Henderson, 1 985;
Zegbe-Dominguez et al. , 2003), decrease fruit size (May and Gonzales, 1 994; Mitchell
et al. , 1 99 1 b), and substantially increase the incidence of fruit rots (May and Gonzales,
1 999; Murray, 1 999). The more severe the irrigation deficit, the greater the loss; May
and Gonzales ( 1 999) suggested that fruit yield from a very early cutoff 80 days
preharvest was 30 Mg ha-1 less than the grower standard; despite improved fruit SSC
with greater water stress, Brix yields (marketable yield x SSC) were significantly lower.
Irrigation cutoffs are also typically irreversible; by the time symptoms of vine-collapse
and dieback are observed (indicators of excessive plant water stress), it is usually too
late to remedy without substantial yield loss. Furthermore, lateral movement of water
can be restricted in fine-textured soils when they are allowed to dry above 50 %
moisture depletion, making it difficult to re-wet the soil in the event that irrigation is
required (Cahn et al. , 2001) . Collectively, these findings make early cutoff approaches
uneconomical.
Irrigation cutoffs implemented within the final 6 weeks preharvest have given
highly variable results, in some fields having no impact on SSC, while in others causing
an unacceptable degree of yield loss. This variability appears largely related to field
specific factors (including soil texture, rooting depth and wetting pattern away from the
drip tape), making it impossible to develop a generic cutoff date recommendation for
drip-irrigated fields. Management tools that help tailor late-season irrigation strategies
to account for field-specific factors are required.
The rapid sensitivity of soil water status to irrigation application, particularly in
the primary root zone, has resulted in soil-based monitoring techniques being favoured
to track plant water availability. Capacitance probes, resistance blocks and tensiometers
have all been investigated (Cahn et ai., 2004; Hanson et al. , 2000a, b ; Hanson and
Peters, 2000). Resistance blocks and tensiometers have been largely favoured to
prevent over- or under-watering, because they are comparatively cheap and can be
automated. However, such measures are not consistently predictive of fruit quality; for
9
example, Cahn et al. (2004) found that fruit SSC could vary as little as 0.4 °Brix over a
soil tension range as large as 225 kPa. This presumably relates to the fact that soil
based measures do not take into consideration other dynamics, such as plant vigour,
disease or cultivar selection.
Other more direct indicators of plant water status including leaf water potential,
infrared leaf thermometry and stomatal conductance exist (Cahn et al. , 2000; Calado et
al. , 1 990; Reid and Renquist, 1 997; Renquist and Reid, 200 1 ; Rudich et aI. , 1 98 1 ;
Zegbe-Dominguez et al. , 2003), although such techniques are often time consuming,
require a high level of user expertise and can be prone to large sampling variability.
Fulton et al. (2001 ) and McCutchan and Shackel (1 992) have reported that covering
plant leaves with a reflective water-impervious bag allowed equilibrium to be reached
between the leaf and stem by effectively stopping transpiration. This method improved
the accuracy of leaf water potential measurements in tree crops by reducing transient
sources of variation, although no work has been published with row crops such as
processing tomato.
In the absence of reliable indicators, many growers have adopted a conservative
approach of full irrigation until cutoff approximately 3 weeks preharvest, choosing to
maximise fruit yield. This approach has been based on that advocated for conventional
irrigation methods. Due to distinctly different soil moisture patterns between furrow
and drip-irrigation techniques, such a late-season cutoff has not consistently increased
fruit SSc. An alternative approach to complete irrigation cutoff is cutback irrigation.
Imposing a controlled water deficit early in the fruit ripening phase has shown the
potential to increase SSC with only a minimal sacrifice in fruit yield (Cahn et al. , 200 1 ;
Renquist and Reid, 200 1 ). Deficit cutback irrigation regimes supply less than 80-90 %
of ETo during fruit ripening. Cahn et al. (2001 ) suggested that at 70 % of ETo, fruit
SSC of an early cutback (initiated when 5- 1 0 % of fruit were ripe) was greater than the
standard grower practice (full cutoff 3 weeks preharvest), with no overall loss in Brix
yield.
Maintaining Brix yield appears to be a useful benchmark for late-season
irrigation management; where Brix yield is significantly reduced, price premiums for
fruit with higher SSC are unlikely to balance the disproportionate loss of yield that has
resulted. Continuing to apply a small amount of irrigation but less than plant demand
may also allow a greater cumulative water stress to be imposed without the risk of
excessive plant canopy dieback. Dieback may increase the incidence of fruit culls,
1 0
which can occur when fruit are directly exposed to solar radiation and tissue over-heats
(Adegoroye and Jolliffe, 1 983). For drip irrigation to continue to gain market
acceptance for processing tomato production reliable strategies are required to manage
late-season irrigation for high fruit SSC and minimal yield loss.
1.4 Statement of Objectives
This study focuses on the potential to improve the management of nutrients and
water for processing tomato crops, with particular emphasis on utilizing the practical
benefits offered by the use of drip irrigation technology. A series of experimental
studies were devised to:
1 .) Evaluate nutrient demand of processing tomato in an ideal root environment
without typical field constraints; uptake determined in this environment was then to be
used as the basis for establishing optimal fertilizer norms and supply patterns in the
field using drip technology.
2 .) Evaluate late-season irrigation management of processing tomato in drip
irrigated fields to increase fruit SSC with minimal yield loss.
1 1
CHAPTER 2: 1998-1999 Aeroponic Nutrition Trial
A glasshouse nutrition experiment was conducted at the Plant Growth Unit
(long. 1 750 37'E; lat. 400 23 'S), Massey University, Palmerston North, during the
summer of 1 998- 1 999. The aim was to evaluate the effect of low and high fertility
during three distinct development phases (vegetative, fruit development and fruit
ripening) on plant growth and subsequent fruit yield and quality of processing tomato.
Aeroponic technology was selected to enable sharp comparison of fertility regimes in
each growth period; supply of nutrients directly at the root surface also effectively
eliminates any buffering capacity of a soil volume, which is a limitation of traditional
fertilizer trials in the field. After determining plant demand under aeroponic conditions,
these results were then to be evaluated in the field using drip fertigation as a method for
improving the efficiency of fertilizer applications by closely matching predicted plant
nutrient demand with supply.
2.1 Materials and Methods
2 . 1 . 1 Experimental S ite
A 9 m x 1 8 m glasshouse was used for this experiment, with an east-west
orientation. A minimum air temperature of 1 5 QC was maintained; side and roof venting
was activated at 25 QC. During the day, relative humidity was between 60-80 %. There
was no enrichment with carbon dioxide. Average outdoor solar radiation during the
summer production months was approximately 20 MJ m-2 day-l (Leathwick et aI. ,
2002).
2 . 1 .2 Production Detai ls
The hybrid cultivar 'Cannery Row' was used in this study as a prominent N.Z.
industry standard. Seed was supplied by Heinz Wattie's Ltd. (Hawkes Bay, N.Z.).
Germinated seedlings were potted into vermiculite in plastic cells (75 ml volume
capacity) 1 0 days after sowing. There was one healthy seedling per cell. Vermiculite
was used because it provides ideal aeration and drainage in hydroponic/aeroponic
culture, and is typically sterile and free from diseases. At field capacity, the physical
1 2
characteristics of the venniculite were 30 % solid, 29 % water-filled pore space and 4 1
% air-filled pore space; total porosity was 70 %. Potted seedlings were raised in a
greenhouse until transplanting into the aeroponic system approximately 5 weeks after
sowing. During this period, all plants were hand-watered daily and fed as required with
a complete fertilizer solution (Pan et aI. , 1 999); of the major nutrients, this solution
supplied approximately 1 00, 30, and 1 00 mg L-) ofN, P and K, respectively. Important
production and phenological dates are summarized in Table 2. 1 .
Table 2.1 . Production and phenological dates, 1998-1999.
Phenological stage
Sown Transplanted First flowering First fruit set First ripe fruit
November 17 December 21 January 20 February 1 March 19
Aeroponic tanks (3.70 m x 1 .25 m x 0.65 m) were made from reinforced
polystyrene lined with black polyethylene. The planting density was initially 40 cm
between double rows and 30 cm within rows. There were 24 harvestable plants per
tank, with an additional two guard plants at each end; guard plants were moved after
each harvest to create equal competition amongst the remaining plants. The final plant
population was a single row with 30 cm in-row spacing; this was equivalent to 22,222
plants ha-), a medium-density field planting (Fig. 2. 1 a, b). At flowering, two boxes of
pollinator bees were placed in the glasshouse; because elevated air temperatures
appeared to reduce bee activity, a hand-held electric vibrator was also used to ensure
adequate pollination and fruit set. Flowers were vibrated three times per week, starting
35 days after transplanting (DAT). Control for white fly (Trialeurodes vaporariorum)
was with a wasp parasitoid (Encarsia formosa). Sticky yellow insect traps were also
placed above plants. No substantial disease pressure was observed.
1 3
Fig. 2.1a. Early-season vegetative growth at 21 OAT.
Fig. 2.1b. Mid-season flowering and fruit development at 42 OAT.
14
2. 1 .3 Experimental Design
Experimental design was randomized complete block with two replications; the
number of replications was constrained by glasshouse size and aeroponic system design.
There were 1 6 aeroponic tanks in total (eight per replicate). Two fertility regimes with
either a low conductivity (L = 0.7 dS m-I) or high conductivity (H = 2.0 dS m-I ) were
delivered factorially to plants during three growth periods, vegetative development
(Pdl ), fruit development (Pd2) and fruit ripening (Pd3). Solution conductivities > 2.5 dS
m-I can limit tomato growth (Maas, 1 986) and were therefore avoided. Because
processing tomato has a detenninate growth habit, each of the three growth periods was
more distinct than in semi-detenninate or indetenninate cultivars (like those grown for
fresh market). Effectively, during vegetative growth there were only two treatments (L
or H). At fruit development these two regimes were split to provide four treatments
(LL, LH, HL or HH), and during fruit ripening they were split again to provide eight
separate treatments (LLL, LLH, LHL, LHH, HLL, HLH, HHL or HHH). A summary
of the factorial treatment structure is provided in Table 2.2. Both the low and high
fertility regimes were based on a complete balanced nutrient solution recommended for
N.Z. greenhouse tomato production (Tregidga et al. , 1 986). A complete description of
both regimes is provided in Appendix 1. All plants received high conductivity (2.0 dS
m-I ) during the initial 1 5 DAT (the aim of the experiment was not to cause extreme
deficiency, as this was not representative of 'nonnal' production practices).
Table 2.2. Factorial treatment design under aeroponics between 1 6-92 OAT.
Vegetative development Fruit development Fruit ripening
1 6-43 OAT 44-64 OAT 65-92 OAT
--c Low (LL) -C Low (LLL)
Low (L) Z High (LLH)
High (LH) -C Low (LHL) High (LHH)
--c Low (HL) -C Low (HLL) High (HLH)
High (H)
-C Low (HHL) High (HH) High (HHH)
4 tanks each 2 tanks each 1 tank each
Z Low fertility (L) was equivalent to 0.7 dS m'1 ; high fertility (H) was equivalent to 2.0 dS m,1
1 5
Inside each aeroponic tank emitters misted treatment solutions directly onto
plant roots for 1 0 seconds every 2 minutes (Fig. 2 .2). Runoff from each treatment
drained directly back into separate 500 L sump reservoirs (in essence, a re-circulating
system). Top-up solutions were 1 00 times the concentration of the standard solution
and were used to adjust reservoir conductivity back to either 0.7 or 2.0 dS m·1 (the low
and high treatments, respectively) . Solution pH was kept at 6.5 using potassium
hydroxide and phosphoric acid. Conductivity and pH were monitored and amended
daily. Reservoir solutions were dumped every 1 0- 1 4 days, reducing the likelihood of
nutrient imbalances (over time, removal of certain nutrients at rates greater than others
may cause solution ratios to change).
Fig. 2.2. Root growth inside the aeroponic tank at 72 OAT.
2. 1 .4 Data Col lection
2 . 1 .4. 1 Leaf Analysis
Representative whole-leaf samples were taken weekly (alongside destructive
growth measurements) to document the N, P and K status of the crop. On each
sampling occasion, 1 0-20 leaves were removed per replicate of each treatment.
Youngest mature leaves (YML; fourth leaf from the plant apex) were selected for
1 6
standardization purposes (Hartz et al. , 1 998). Leaves were oven dried at 65 °C for 48
hours and then ground to pass through a 0.5 mm screen. Total N and P concentrations
(expressed as % of dry mass) were determined colourimetrically following Kjeldahl
acid digestion (Twine and Williams, 1 97 1 ). Total K concentration (expressed as % of
dry mass) was determined by flame-emission spectroscopy following nitric acid
digestion (Technicon, 1 973).
2 . 1 .4.2 Destructive Whole-Plant Sampling
There were 1 2 destructive plant harvests between 8 and 92 DAT; a summary of
harvest dates relative to treatment regimes is provided in Table 2.3 . At each harvest
two plants were removed from each aeroponic tank. During the phase of establishment
when only high fertility was supplied (0- 1 5 DAT), 1 6 plants per replicate were therefore
sampled (all 8 tanks in the replicate x 2 plants). However, in each subsequent period
the number of plants sampled decreased due to the factorial design. During vegetative
development, sampling of each treatment was equal to 8 plants per replicate (4 tanks x 2
plants). During fruit development, sampling of each treatment was equal to 4 plants per
replicate (2 tanks x 2 plants). During fruit ripening, sampling of each treatment was
equal to 2 plants per replicate ( 1 tank x 2 plants).
On each sampling occasion, the fresh biomass of individual whole-plants was
recorded. Plants were then separated into leaf, stem, fruit and root tissues, and each
component weighed. A representative leaf sample was further separated into leaf
lamina and petiole material and the mass of each recorded. Leaf laminae were used to
measure leaf area by leaf area meter (Model 3 1 00, LI-COR Biosciences, Lincoln, NE).
A representative stem sample was also collected and weighed. Samples were oven
dried at 65 °C for 48 hours before being weighed dry and ground to pass a 0.5 mm
screen. The entire root biomass was dried, weighed and ground. The handling of fruit
tissue is described separately as yield and quality (Section 2 . 1 .4.3). Ground leaf and
stem samples were analyzed for Total N, P and K concentration as previously
described. Nutrient concentrations could not be successfully determined for root tissue
due to what appeared to be silicon contamination in the ground samples (from dried
vermiculite that was inadvertently incorporated with roots).
1 7
Table 2.3. Destructive plant sampling between 8-92 OAT relative to fertility treatments.
Harvest ocassion Sampling date OAT
All plants received H during first 15 days Z 1 December 29 8 2 January 5 1 5
Treatments L and H initiated at 1 6 days
3 January 12 22 4 January 1 9 29 5 January 26 36 6 February 2 43
Change over from L and H to LL, LH, HL and HH at 44 days
7 February 9 50 8 February 1 6 57 9 February 23 64
Change over from LL, LH, HL, HH to LLL, LLH, LHL, LHH, HLL, HLH, HHL, HHH at 65 days
1 0 March 2 71 1 1 March 1 1 80 1 2 March 23 92
Z Low fertility (L) was equivalent to 0.7 dS m-1 ; high fertility (H) was equivalent to 2.0 dS m-1
2 . 1 .4.3 Fruit Yield and Quality Measurements
Of the twelve whole-plant harvests made, only eight resulted in a measurable
fruit biomass. Beginning 36 DAT, all fruit were hand-harvested and weighed to
determine the total yield of each treatment. Fruit were sorted according to four
categories, including marketable (sound, intact red and coloured fruit), green (no
colour), blossom-end rot fruit (BER) and pathological rots. The number and mass of
fruit in each category was recorded, and mean individual fruit mass calculated. At each
harvest, approximately 20-25 intact fruit per replicate of each treatment (representing
the maturity at that time) were selected for nutrient analysis; fruit that had lost integrity
due to either BER or pathological rot were not included in these analyses. Fruit were
cut longitudinally, weighed fresh and then oven dried at 65 °C before being weighed
dry. Samples were subsequently ground to pass through a 0.5 mm screen and analyzed
for Total N, P and K concentration as previously described. Additionally, the
concentration of Total Ca and Mg (expressed as % of dry mass) were determined by
atomic absorption spectroscopy following nitric acid digestion (Technicon, 1 973) .
At the final harvest (92 DAT), 20-25 intact marketable fruit per replicate of each
treatment were randomly collected to assess quality. Fruit were mechanically juiced,
1 8
and a filtered sample (using cheese-cloth) was collected for determination of SSC and
titratable acidity. Fruit SSC (OBrix) was measured using a temperature-compensated
refractometer (NI , Atago Ltd. , Tokyo, Japan). Titratable acidity was measured using an
automatic pH titrator (DL 2 1 , Mettler-Toledo Inc. , Columbus, OH). To determine
titratable acidity, 1 mL of juice was diluted to 50 mL in distilled water. With
continuous stirring, the sample was titrated against 0. 1 N sodium hydroxide to pH 8 . 1
(Mencarelli and Saltveit, 1 988). Results were expressed as the equivalent mass o f citric
acid as a percent of total dry mass; citric acid is the predominant organic acid in ripe
tomatoes (Davies and Hobson, 1 98 1 ). Lycopene content was determined on the
unfiltered pulp sample. From the blended sample, 1 00 ).lg of tomato pulp was added to
7.0 mL of 4:3 (v/v) ethanol:hexane mixture (Barrett and Anthon, 200 1 ). Samples were
vortexed intermittently for one hour out of bright light, after which time 1 .0 mL of
distilled water was added to each sample and briefly mixed. Samples were left to stand
for 1 0 minutes allowing the solvent phases to separate. The hexane layer was removed
and the absorbance read by spectrophotometer (Model U-2000, Hitachi Ltd. , Tokyo,
Japan) at 503 nm. At 92 DAT, harvest index (H.I.) was determined as the proportion of
fruit yield to total plant biomass (both dry).
2 . 1 .4.4 Growth Analysis
The primary data used in growth analysis calculations were leaf area and dry
plant biomass (leaf and total). Total plant biomass referred only to the above-ground
constituents (leaf, stem and fruit) ; root biomass was excluded to allow comparisons
with previous studies (in the field, accurately exhuming roots has proven difficult).
Growth analysis variables were calculated for each defined period of plant
development; because fertility regime targeted three distinct phases of growth, this
approach offered the greatest practical insight into overall treatment effects. The
following formulas were used:
Where:
2. 1 .4.4a RELATIVE GROWTH RA TE (RGR)
RGR = Log. T dwZ - Log. T dwl / tz - tl
RGR = relative growth rate (g g-l dail)
1 9
Where:
Where:
Where:
Where:
Tdw = total plant dry biomass (g)
= time (days)
1 . 2 = subscripts denote previous and current harvests, respectively
2. I .4.4b NET ASSIMILATION RATE (NAR)
NAR = net assimilation rate (g cm-2 day-I)
T dw = total plant dry biomass (g)
LA = leaf area (cm2)
= time (days)
1 . 2 = subscripts denote previous and current harvests, respectively
2. I .4.4c LEAF AREA RATIO (LAR)
LAR = LA I Tdw
LAR = leaf area ratio (cm2 g"1 )
LA = leaf area (cm2)
T dw = total plant dry biomass (g)
2. I .4.4d LEAF WEIGHT RATIO (L WR)
L WR = Ldw I T dw
L WR = leaf weight ratio
Ldw = leaf dry biomass (g)
T dw = total plant dry biomass (g)
2. 1.4.4e SPECIFIC LEAF AREA (SLA)
SLA = LA / Ldw
SLA = specific leaf area (cm2 g"l Ldw) LA = leaf area (cm2)
Ldw = leaf dry biomass (g)
20
2. 1 .4.5 Nutrient Uptake
Nutrient uptake into leaf, stem and fruit tissue was calculated for N, P and K;
total uptake of each nutrient was the sum of these constituents. Because fertility
regimes targeted vegetative development, fruit development and fruit ripening,
cumulative uptake of N, P and K was calculated at the end of each period of growth.
The following formulas were used:
Where:
Where:
Where:
2. 1 .4.5a LEAF UPTAKE
Leaf uptake = Ldw x % N, P or K in leaf
Ldw = leaf dry biomass (g planrl)
% N, P or K = concentration of nutrient in the leaf sample
2. 1 .4.5b STEM UPTAKE
Stem uptake = Sdw x % N, P or K in stem
Sdw = stem dry biomass (g plan(1 )
% N, P or K = concentration of nutrient in the stem sample
2. 1 .4.5c FRUIT UPTAKE
Fruit uptake = Fdw x % N, P or K in fruit
Fdw % N, P or K
= fruit dry biomass (g planrl)
= concentration of nutrient in the fruit sample
2 . 1 .5 Statistical Analysis
Experimental data were subjected to ANOVA tests usmg the SAS General
Linear Model (GLM) procedure (SAS Institute Inc. , Cary, N.C.). Means separation was
made on significant ANOV A tests using the Least Significant Difference (LSD) method
(p < 0.05). Where relevant, statistical analyses were made on log-transformed (to
render variability more homogenous with time) or arc-sine-transformed {to ensure
21
normality assumptions were met) data sets. Because ANOV A outcomes were
consistently poor (clearly limited by replication constraints) a series of orthogonal
contrasts were made as a means to compare specific treatments (Table 2 .4); such
contrasts allowed combinations of interest to be considered within the overall analysis
(therefore not losing degrees of freedom). This technique generally clarified plant
responses to fertility, which were often numerically distinct between fertility regimes .
Most contrasts focused on early-season fertility, primarily due to the determinate
growth habit in processing tomatoes. Periods of vegetative and fruit development
typically account for most active plant growth (during ripening it is predominantly
metabolic changes that occur in fruit). Therefore, many plant responses appeared likely
to be influenced by fertility during the initial two periods of growth. Additional
contrasts were made where the expected response to fertility warranted (i.e. those
relating to fruit quality, which may be influenced by late-season fertility).
SAS non-linear (NUN) regression procedures were used to test the significance
of logistic modelling in describing seasonal trends in biomass accumulation.
Correlation analysis was used to report positive and negative linear relationships
between variables of interest.
Table 2 .4. Standard orthogonal contrasts made during each period of plant growth.
Growth period
Vegetative development Z
Fruit development
Fruit ripening
Contrast description Abbreviation
LL vs. H H border treatments (LL + LH) vs. ( HL + HH) Lx vs. Hx
(LLL + LLH) vs. (LHL + LHH) LLx vs. LHx (HLL + HLH) vs. (HHL + HHH) HLx vs. HHx
LLL vs. HHH border treatments (LLL + LLH + LHL + LHH) vs. (HLL + HLH + HHL + HHH) Lxx vs. Hxx
Z No contrasts were made during the initial period because there were only two treatments
22
2.2 Results
2.2. 1 Fertil ity Index
Field studies typically involve a known fertilizer application (e.g. kg ha- I ) ;
however, there was no practical method in which to quantify the mass of nutrients that
were applied seasonally using the aeroponic system. Calculated nutrient uptake was
only indicative of that absorbed by plants for growth; this was not equivalent to the
mass of nutrients 'applied' . It would also have been misleading to represent the
quantity of fertilizers placed in the starter and top-up solutions as the quantity of
nutrients applied. This difficulty was not unique to aeroponic study; most, if not all ,
solution-culture experiments are affected in a similar manner. Nevertheless, the
aeroponic approach remained a convenient and effective way to study growth responses
to nutrient supply while controlling other potentially limiting factors. The design of the
experiment was in fact such that either sub-optimal or optimal fertility (low and high
treatments, respectively) were applied, where the quantity of each mineral was less
important than availability.
As an alternative to the mass of nutrients applied, a weighted fertility index (FIN)
was created allowing crop nutrition to be correlated to variables of interest. The index
was calculated for each of the eight treatment combinations by calculating mean
solution conductivity across the three defined growth periods (Eq. 2. 1 ); these values
therefore fell between 0.7 and 2.0 dS m-I , the range of the two border treatments
receiving either continuous low or high fertility (LLL and HHH, respectively; Table
2.5).
Where:
FIN = {NI X EC during Pdd + (Nl x EC during P�) + eN] x EC during Pd31 (NI + N2 + N3)
NI, N2 and N3 = the number of days during Pd" Pd2 and Pd3, respectively
EC (dS m-I) = low (0.7) or high (2.0) conductivity during Pd" Pd2 and Pd3
(Eq. 2. 1)
Although the index did not directly weight the relative importance in the timing
of these regimes, a poor correlation between FIN and the response variable may suggest
that timing of fertility was more important than the mean regime concentration across
all periods. An adjusted fertility index (FINAdj.) was also created for treatment
23
combinations across only the final two growth periods (Eq. 2.2); this index was
specifically designed for correlation with fruit quality variables, which are typically
determined after fruit have set. As with the original index, these values continued to
fall between 0.7 and 2.0 dS m-I , the range of the LLL and HHH treatments, respectively
(Table 2.5).
Where:
FINAdj. = (N� x EC during Pg) + (Nl x EC during Pd31
(Nz + N3)
N2 and N3 EC (dS m-I)
= the number of days during P d2 and P d3, respectively
= low (0.7) or high (2.0) conductivity during P d2 and P d3
(Eq. 2.2)
Table 2.5. Fertility regime, fertility index (FIN) and adjusted fertility index (FINAdjJ between
1 6-92 OAT.
Fertility regime (dS mol) FIN FINAdj.
Treatment schedule Z Pd, Pd2 Pd3 (1 6-43 OAT) (44-64 OAT) (65-92 OAT) (dS m-' ) (dS mol )
LLL 0.7 0.7 0.7 0.7 0.7 LLH 0.7 0.7 2.0 1 .2 1 .4 LHL 0.7 2.0 0 .7 1 . 1 1 . 3 LHH 0.7 2.0 2.0 1 .5 2 . 0 HLL 2.0 0.7 0 .7 1 .2 0 .7 HLH 2.0 0.7 2.0 1 .6 1 . 4 HHL 2.0 2.0 0.7 1 .5 1 . 3 HHH 2.0 2.0 2.0 2.0 2.0
Z Low fertility (L) was equivalent to 0.7 dS mol ; high fertility (H) was eqivalent to 2.0 dS m-l
2.2.2 Destructive Whole-Plant Sampling
Twelve destructive growth measurements were made between 8 and 92 DA T.
An error in biomass measurements made 7 1 DAT was found; all data collected from
this date were subsequently omitted.
2.2.2 . 1 Total Plant Dry Biomass
Most field studies have been unable to accurately exhume the below-ground
portion of the tomato plant; using aeroponics root biomass was easily retrieved.
However, to allow comparisons with previous experiments root biomass was considered
24
independently to above-ground constituents. Total plant dry biomass (Tdw) therefore
only comprised leaf, stem and fruit tissue. Fertility regime during P d) appeared to
largely determine seasonal Tdw accumulation (vegetative framework was established
during this period, as was flower number and therefore subsequent fruit number).
Accordingly, to display seasonal accumulation of plant biomass, treatments were
grouped according to the low or high fertility regime received during this initial period
(designated Lxx and Hxx, respectively; Fig. 2 .3); Lxx included the LLL, LLH, LHL and
LHH treatments, whereas Hxx included the HLL, HLH, HHL and HHH treatments. To
further illustrate the close relationship between biomass productivity and initial fertility
regime, the continuous low and high border treatments (LLL and HHH, respectively)
were also retained for comparison against the Lxx and Hxx groupings.
-' ..... c ca 0.. 0) ....-
� "C I-
500.--------------------------------------------------,
400
300
200
1 00
. . . . . . . . . . . . �dl . . . I · · · · · Pd2 . . . . . . 1 . . o
o 1 0 20 30 40 50 60
Days after transplanting
Pd3 . . . . . .
70 80
. . . . . . . . .
90
1
100
Fig. 2.3. Effect of fertility regime on seasonal total plant dry biomass (Tdw) accumulation. Treatment means are given ± standard error (for Lxx and Hxx only). Treatments were T dw for Lxx (---\7---), Hxx (---0---), LLL (-.... -) and HHH (-e-). Tdw included leaf, stem and fruit dry biomass. Significant logistic growth models were fitted to each treatment (p < 0.05).
Total dry weight accumulated steadily during P d ) in both treatments, largely
attributed to vegetative (leaf and stem) growth. There was an increasing numerical
trend towards greater Tdw in plants receiving high fertility during this initial period,
although at 43 DAT this effect was not significant (Table 2.6). During Pd2, Tdw
accumulated rapidly in all treatments, largely attributed to fruit growth. The difference
25
between the Lx (LL and LH) and Hx (HL and HR) treatments increased during this
phase of growth, although the overall effect of fertility regime was not significant at 64
DAT. Contrasts on specific treatment means were also not significant at this time,
despite a continued trend towards greater Tdw with higher fertility.
Table 2.6. Effect of fertility regime on total above-ground dry biomass (Tdw) accumulation
at the end of each defined growth period.
Fertility
regime
LLL LLH LHL LHH HLL HLH HHL HHH
SE p -value
Contrasts LL vs. HH Lx vs. Hx
LLx vs. LHx H Lx vs. HHx LLL vs. HHH Lxx vs. Hxx
Total above-ground dry biomass (g)
43 OAT (end of Pd1 )
1 1 4
124
9 ns
64 OAT (end of Pd2)
253
278
286
302
1 8 ns
ns ns
92 OAT (end of Pd3)
314 361 288 302 433 414 385 431
1 8 ns
ns ns
�**
ns , * , *** non-significant at p < 0 . 10, or significant at p < 0.10 or 0.01 , respectively
The rate of Tdw accumulation slowed during Pd3, a pattern consistent with fruit
ripening. This slowing was more distinct in plants that had initially received low
fertility. At 92 DAT there was no difference in Tdw between the LLL and grouped Lxx treatments. By comparison, Tdw gain was greater in treatments that had received high
fertility during P d! ; however, there was also little difference between the HHH and
grouped Hxx treatments at 92 DA T. Despite large numerical differences in plant
biomass between fertility regimes, the overall effect of treatment on T dw was not
significant at 92 DAT. Statistical separation was clearly limited by the number of
replicates possible using the glasshouse setup. Contrasts clarified the T dw response to
fertility at 92 DAT; both HHH and Hxx fertility regimes resulted in significantly greater
biomass accumulation when compared separately to LLL and Lxx fertility
(respectively) . Total plant dry biomass at 92 DAT was weakly correlated with FIN (P <
26
0.08, r = 0.46). This relationship appeared overstated, because LLL fertility achieved
similar Tdw to LHH fertility, as was also the case for HLL and HHH fertility; these
treatment comparisons represented the most dissimilar regimes after the initial period
(i.e. those most likely to be different if FIN across all periods was important). These
observations instead implied the importance of adequate early-season fertility in
determining plant biomass outcomes.
2.2.2. 1 a LEAF BIOM4SS
Leaf dry biomass (Ldw) included both leaf lamina and petiole material. Seasonal
Ldw accumulation was sigmoid, distinguished by a large plateau after fruit began to
develop; because Ldw was largely detennined in response to fertility during Pdl , only the
grouped Lxx and Hxx treatments were retained figuratively (Fig. 2.4). Similar patterns
were also observed for the continuous low and high border treatments.
1 40
1 20
1 00 -
'- 80 -c: t1l a. 0> 60 -........-
� '" ...J 40 -
20 -
0 - . . . . . . . . . . 1 . . . . Pd2 · · · · · · · · 1 · · I I I I I I I I I
0 1 0 20 30 40 50 60 70 80 90 1 00
Days after transplanting
Fig. 2.4. Effect of fertility regime on seasonal leaf dry biomass (Ldw) accumulation. Treatment means are g iven ± standard error. Treatments were Ldw for Lxx (---\7---) and Hxx (---0---). Ldw included both leaf lamina and petiole dry biomass. Significant logistic growth models were fitted to both treatments (p < 0.05).
Leaf dry weight accumulated rapidly in both treatments during Pdl . Leaf
biomass was uncharacteristically low at 36 DAT, but appeared unrelated to fertility
27
regime; no noticeable decline in Tdw was observed on this date. There was an
increasing numerical trend towards greater Ldw in plants receiving high fertility during
P d) , although this effect was not significant at 43 DAT (Table 2.7). Plants initially
receiving low fertility had accumulated most leaf biomass by 43 DAT. By comparison,
plants initially receiving high fertility continued to accumulate new leaf biomass into
the early part of P d2; however, there was no difference in Ldw between the HL and HH
treatments, suggesting this dynamic was largely controlled by high fertility during the
initial period. Despite numerical differences between the Lx and Hx treatments, the
effect of fertility on Ldw accumulation remained not significant at 64 DAT. The
coefficient of variation (cv, a measure of sampling variability) was large on this
occasion (cv = 43 %). The lack of statistical separation at 64 DAT appeared a transient
effect related to this variability; on the two previous harvest occasions during Pd2 (50
and 57 DAT), Hx fertility resulted in significantly greater leaf biomass than Lx fertility
(p < 0.06 and 0.07, respectively).
Table 2.7. Effect of fertility regime on leaf dry biomass (Ldw) accumulation at the end of
each defined growth period.
Fertility
regime
LLL
LLH
LHL
LHH
HLL
HLH
HHL
HHH
SE p -value
Contrasts LL vs. HH
Lx vs. Hx
LLx vs. LHx
H Lx vs. HHx
LLL vs. HHH
Lxx vs. Hxx
ns non-significant at p < 0 . 10
43 DAT
(end of Pd1 )
62
74
6 ns
Leaf dry biomass (g)
64 OAT 92 OAT
(end of Pd2) (end of Pd3)
73 73 80
83 39 49
10 1 1 06 72
1 02 60 95
1 0 9 ns ns
ns ns
ns ns ns ns
During P d3 there was effectively no new accumulation of Ldw in any treatment;
this was consistent with the determinate growth habit in processing tomatoes, where
28
new vegetative growth stops once fruit sinks have developed. Although Ldw in the Lxx
and Hxx treatments declined late in the season, this observation appeared to simply
reflect plant-to-plant differences (sample size decreased as the experiment progressed)
rather than a specific growth response to fertility; leaf die-back in the glasshouse
environment was minimal. Furthermore, leaf biomass for the LLL and HHH border
treatments remained largely unchanged during this period (as representative of the
fertility 'extremes'). Despite distinct numerical differences between the Lxx and Hxx regimes at 92 DAT, the overall effect of fertility on Ldw was not significant. Contrasts
at this time were also not significant. Presumably this reflected the high sampling
variability observed at 92 DAT (cv = 54 %). Leaf biomass dynamics so late in the
season were of little practical relevance to yield outcomes, because fruit number and
size were already determined. There was no correlation between Ldw at 92 DAT and
F IN, confirming that the timing of low or high fertility was more important than the
mean regime concentration across all periods.
2.2.2. 1h STEM BIOMASS
Stem dry biomass (Sdw) included all above-ground tissue that was not otherwise
leaf or fruit. Seasonal Sdw accumulation was sigmoid, also characterized by an
increasing plateau as fruit developed (particularly under Lxx fertility) . Although Sdw
accumulation did not truncate distinctly at the end of P dI , the stem framework still
appeared to be largely a function of fertility during this initial phase of plant growth; for
continuity, the grouped Lxx and Hxx treatments were retained figuratively (Fig. 2.5) .
Similar patterns were also observed for the continuous low and high border treatments.
Stem biomass accumulated steadily in both treatments during PdI . At 36 DAT,
Sdw was comparatively high, although was matched by a small decline in Ldw as
previously noted. The effect of fertility regime on Sdw was not significant at 43 DA T
(Table 2.8). After 57 DAT there was very little new stem biomass developed in plants
that had initially received low fertility (at 80 DAT, stem biomass was abnormally high).
By comparison, plants initially receiving high fertility continued to accumulate Sdw late
into the season. Despite distinct numerical differences between the Lx and Hx regimes,
the overall effect of fertility was not significant at 64 DAT. However, contrasts on
specific treatment means were significant at this time; both HH and Hx fertility resulted
in greater Sdw accumulation than LL and Lx fertility, respectively.
29
1 40
1 20 -
1 00 -
....... 80 -c ro 0..
0> 60 -.......... � "
Cf) 40
20
0
0
.. "' ....... � � 1
I 1 0
I 20
. . . . . . . . . . . . .. 1 . . . . . Pd2 . . . . . . . . .. 1 . . . . . · · · · · · · · · 1 I I I I I I I
30 40 50 60 70 80 90 1 00
Days after transplanting
Fig. 2.5. Effect of fertility regime on seasonal stem dry biomass (Sdw) accumulation. Treatment means are given ± standard error. Treatments were Sdw for Lxx (---\7---) and Hxx (---0---). Sdw included all above-ground tissue that was not otherwise considered as leaf or fruit. Significant logistic growth models were fitted to both treatments (p < 0.05).
Similar statistical trends were also established at 92 DAT; HHH and Hx.x
fertility resulted in greater Sdw accumulation than LLL and Lxx fertility, respectively.
The detailed means separation at 92 DAT appeared to be of limited practical relevance,
because fruit number and size were already determined; additionally, both of these yield
components relate more closely to assimilate supply from leaves rather than the stem.
Although Sdw at 92 DAT was weakly correlated with FIN (P < 0.06, r = 0.59), this
appeared to be an anomaly in light of the contrasts made (LLx vs. LHx and HLx vs.
HHx were both not significantly different).
30
Table 2.8. Effect of fertility regime on stem dry biomass (Sdw) accumulation at the end of each defined growth period.
Fertility
regime
LLL LLH LHL LHH HLL HLH HHL HHH
SE p -value
Contrasts LL vs. HH Lx vs. Hx
LLx vs. LHx HLx vs. HHx LLL vs. HHH Lxx vs. Hxx
43 OAT
(end of Pd1 )
44
42
4 ns
Stem dry biomass (g)
64 OAT
(end of Pd2)
59
71
92
93
7 ns
** **
92 OAT
(end of Pd3)
59 d 84 bc 60 d 80 c
1 1 2 a 1 35 a 1 1 1 ab 1 02 abc
7
ns ns ***
ns , **, *** non-significant at p < 0. 1 0, or significant at p < 0.05 or 0.01 , respectively; means withi n columns separated using the Least Significant Difference (LSD) test, p < 0.05
2.2.2. 1c FRUIT BIOMASS
Fruit dry biomass (Fdw) included all fruit that had a diameter greater than 1 0
mm. Seasonal Fdw accumulation was generally sigmoid. Because total fruit number
and total yield at 92 DAT appeared largely detennined according to plant growth during
Pdl (fruit yield dynamics are described in Section 2.2 .4) the grouped Lxx and Hxx treatments were retained figuratively (Fig. 2.6); almost identical patterns were observed
for the continuous low and high border treatments.
A measurable fruit mass did not occur until 36 DAT; because only fruit of a
certain size threshold were collected, fruit biomass may have been slightly
underestimated on earlier dates. There was no effect of fertility regime on Fdw during
Pdl , with equivalent fruit biomass at 43 DAT (Table 2.9). During Pd2, Fdw accumulated
rapidly in all treatments; fruit biomass was numerically greater in plants that had
initially received low fertility, which may have reflected slightly earlier flowering.
However, this difference in fruit biomass was not significant at 64 DAT. Contrasts on
specific treatment means at this time were also not significant.
3 1
280
240
200
'...- 1 60 c:: ca a. C> 1 20 .......
� ." LL
80
40
0 -I
0 1 0 20 30 40 50 60 70 80 90 1 00
Days after transplanting
Fig. 2.6. Effect of fertility regime on seasonal fruit dry biomass (Fdw) accumulation. Treatment means are given ± standard error. Treatments were Fdw for Lxx (---\7---) and Hxx (---0--). Fdw included all fruit with a diameter greater than 1 0 mm. Significant logistic growth models were fitted to both treatments (p < 0.01) .
Accumulation of Fdw in plants initially receiving low fertility slowed during Pd3 ;
by comparison, plants initially receiving high fertility continued to accumulate Fdw.
This phase of extra biomass accumulation largely accounted for the difference in F dw
between Lxx and Hxx treatments. Despite quite large differences, the overall effect of
fertility regime on Fdw was not significant at 92 DAT. However, contrasts at this time
clarified the general effect of fertility on fruit biomass; both HHH and Hxx fertility
regimes resulted in significantly greater Fdw accumulation than LLL and Lxx fertility,
respectively. The remaining contrasts were not significant. Fruit dry biomass at 92
DAT was weakly correlated with FIN (P < 0.09, r = 0.44). As with previous biomass
measures, this relationship appeared overstated because LLL fertility achieved similar
Fdw to LHH fertility, as was also the case for HLL and HHH fertility. Collectively,
these observations reinforced that the timing of low or high fertility was most important
in determining yield outcomes rather than the mean regime concentration across all
periods. The effect of fertility regime on harvest index (H.!.) was not significant at 92
DAT. Across treatments, the index averaged 0.56; a typical H.!. for processing tomato
grown in the field is approximately 0.60 (T.K. Hartz, personal communication).
32
Table 2.9. Effect of fertility regime on fruit dry biomass (Fdw) accumulation at the end of
each defined growth period.
Fertility
regime
LLL LLH LHL LHH HLL HLH HHL HHH
SE p -value
Contrasts LL vs. HH Lx vs. Hx
LLx vs. LHx HLx vs. HHx LLL vs. HHH Lxx vs. Hxx
43 OAT
(end of Pd1 )
8
8
ns
Fruit dry biomass (g)
64 OAT
(end of Pd2)
121
124
93
106
9 ns
ns ns
92 OAT
(end of Pd3)
1 82 1 96 1 88 1 73 2 1 5 208 2 1 4 234
7 ns
ns ns
** ns , *, ** non-significant at p < 0.10 , or significant at p < 0. 1 0 or 0.05, respectively
2.2.2. 1d ROOT BJOMASS
Seasonal root dry biomass (RIw) accumulation was sigmoid, distinguished by a
large plateau after fruit development began. Although RIw accumulation did not
truncate distinctly at the end of P d I , most of root framework was established during this
initial phase of growth; for continuity, the grouped Lxx and Hxx treatments were
retained figuratively (Fig. 2 .7). Similar patterns were also observed for the continuous
low and high border treatments.
Root biomass accumulated rapidly in both treatments during Pdl . The small
trend towards greater RIw in plants initially receiving high fertility was not significant at
43 DAT (Table 2. 1 0). Root biomass was comparatively low at 43 DAT and high at 5 0
DAT; neither observation appeared related to fertility regime. Both dates fell during the
phase in which root growth slowed substantially. Despite continued numerical
differences between the Lx and Hx treatments, the overall effect of fertility was not
significant at 64 DAT. The contrast between the two border regimes was however
significant, with greater RIw under higher fertility.
33
40,-------------------------------------------------�
35 -
30
25
'-� 20 a.
1 5
a:." 1 0
5
o
o
:t-: ;.-:: "
:::: :>. � I · · · · ·
1 0 20
I I - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -1-
_-/-r- --- -� --l--- - - - -- - - - - - - -I--- - - - - - - - -" l --- -
1/J/o /
" I ,J;'
�91 . . . . . . . . . . . . . 1 . . . . . . . . . . . . Pd2 . . . . . . 1 . . I I
30 40 50 60 70
Days after transplanting
. . �d� . . . . . . . . . . . . . . 1 I I
80 90 1 00
Fig. 2.7. Effect of fertility regime on seasonal root dry biomass (�w) accumulation. Treatment means are given ± standard error. Treatments were �w for Lxx (---'1---) and Hxx (---0---). Significant logistic growth models were fitted to both treatments (p < 0.05).
There was little new accumulation of Raw during P d3 . Despite the small trend
towards greater Raw with high initial fertility, this effect was not significant at 92 DAT;
across treatments, root biomass averaged approximately 7 % of whole-plant biomass at
this time. The significance of the HLx vs. HHx contrast appeared overstated due to low
and high Raw in the HHL and HLH treatments, respectively. Such changes so late in the
season were uncharacteristic of determinate tomatoes grown in a disease-free root
environment. Additionally, Raw did not change greatly in the two border treatments
during this period (as representative of fertility 'extremes'). All other contrasts were
not significant. As expected, Raw at 92 DAT was not significantly correlated with F IN.
34
Table 2.10. Effect of fertility regime on root dry biomass (�w) accumulation at the end of
each defined growth period.
Fertility
regime
LLL LLH LHL LHH HLL HLH HHL HHH
SE p -value
Contrasts LL vs. HH Lx vs. Hx
LLx vs. LHx HLx vs. HHx LLL vs. HHH Lxx vs. Hxx
43 OAT
(end of Pd1 )
1 5
18
ns
Root dry biomass (g)
64 OAT
(end of Pd2)
1 9
23
20
27
ns
ns
92 OAT
(end of Pd3)
21 31 29 23
30 36 21 28
2 ns
ns
ns ns
ns , *, ** non-significant at p < 0 . 10 , or Significant at p < 0. 1 0 or 0.05, respectively
2.2.2.2 Leaf Area Measurements
Seasonal leaf area (LA) expansion was sigmoid, distinguished by a large plateau
as fruit developed. There was a strong correlation between LA and Ldw (P < 0.0 1 , r =
0.92). Because leaf dynamics were largely determined during the initial period of plant
growth, only the grouped Lxx and Hxx treatments were retained figuratively (Fig. 2.8).
Similar patterns were also observed for the continuous low and high border treatments.
35
20000
1 6000
� '� 1 2000 a. N E o � 8000
4000
0 -
o
� " " 'f" "
1 0
/ 1 1 1 1 /1:/ /!'--' -
:0: ,/ ,,' �/ ,1 , , , , , , , , , ,
� /" ;'-;�;
20
· · · · · · · 1 Pd2 . . . . . . . . . . . . I · · I I I I
30 40 50 60 70
Days after transplanting
I 80 90 1 00
Fig. 2.8. Effect of fertility regime on seasonal leaf area (LA) expansion. Treatment means are given ± standard error. Treatments were LA for Lxx (---"7---) and Hxx (---0---). LA included leaf lamina only. Significant logistic growth models were fitted to both treatments (p < 0.05).
Leaf area expanded rapidly in both treatments during Pd1 . Measurements were
uncharacteristically low at 36 DAT, although consistent with low Ldw also observed;
these dynamics appeared unrelated to fertility treatments. The trend towards greater LA
in plants receiving high initial fertility was significant at 43 DAT (Table 2 . 1 1 ). After 50
DAT, there was essentially no new leaf area expansion under any treatment regime.
Despite larger LA with high initial fertility, the overall effect of treatment was not
significant at either 64 or 92 DA T. Contrasts on specific treatment means were also not
significant. Sampling variability was large on both occasions (cv = 42 and 49 %,
respectively), limiting statistical power. As previously mentioned, this was due to
plant-to-plant variations rather than a large incidence of leaf die-back. Leaf area at 92
DAT was not significantly correlated with FIN. As with Ldw, these observations
confirmed that the timing of fertility was more important to LA development than the
mean concentration across all periods of growth.
36
Table 2.1 1 . Effect of fertility regime on leaf area (LA) accumulation at the end of each
defined growth period.
Fertility
regime
LLL LLH LHL LHH HLL HLH HHL HHH
SE p -value
Contrasts LL vs. HH Lx vs. Hx
LLx vs. LHx HLx vs. HHx LLL vs. HHH Lxx vs. Hxx
43 OAT
(end of Pd1 )
8928
1 3064
1 292
Leaf area (cm2)
64 OAT
(end of Pd2)
1 3742
1 3338
1 9 1 56
1 6960
1 826
ns
ns ns
ns • • non-significant at p < 0 . 10 . or significant at p < 0 . 10 . respectively
2.2.2.3 Plant Growth Analysis Measurements
2.2.2.3a RELATIVE GROWTH RATE
92 OAT
(end of Pd3)
1 3728
1 2390 751 9
9252
20164
1 5009
1 2727 1 5478
1 470 ns
ns ns ns ns
The relative growth rate (RGR) of plants declined seasonally in all treatments
(Table 2 . 1 2). Across periods, RGR was numerically greater under Hxx fertility
compared to Lxx fertility; although this difference was small, it was nevertheless
consistent with total biomass accumulation (which was also greater with high initial
fertility). Despite this trend, the effect of treatment on RGR was not significant for any
period. As with Tdw, this observation appeared largely related to insufficient
replication. Contrasts on specific treatment means were also not significant. Growth
rates were increasingly variable late in the season; this appeared to reflect observed
plant-to-plant variations (particularly in vegetative biomass) rather than important
growth responses. Across periods, reinstating high fertility after a phase of low fertility
had no consistent effect on RGR.
37
Table 2 .12 . Effect of fertility regime on relative growth rate (RGR) of plants across each
defined period of g rowth.
Fertility
regime
LLL LLH LHL LHH HLL HLH HHL HHH
SE p -value
Contrasts LL vs. HH Lx vs. Hx
LLx vs. LHx H Lx vs. HHx LLL vs. HHH Lxx vs. Hxx
ns non-significant at p < 0 .10
Relative growth rate (g g.1 week·1)
Pd2 Pd3 Pd1 (1 6-43 OAT) (44-64 DAT) (65-92 OAT)
0 .76
0 .78
0.026
ns
0.26
0.29
0.28
0.30
0.030
ns
ns ns
0.06
0 .09
0.01
0 .03
0 . 1 0 0 .09
0 .06
0 .09
0.0 1 7
ns
ns ns ns ns
2.2.2.3b NET ASSIMILA TION RATE AND LEAF AREA RATIO
The net assimilation rate (NAR) of plants declined seasonally in all treatments
(Table 2 . 1 3) . The effect of fertility regime on NAR was not significant, suggesting the
capacity of each leaf unit to generate new biomass was similar; across treatments, NAR
averaged 0.007, 0.004 and 0.002 g cm-2 week- 1 during Pdl , Pd2 and Pd3, respectively.
Contrasts on specific treatment means were also not significant.
Leaf area ratio (LAR) also declined seasonally in all treatments (Table 2 . 13) .
This reflected the increasing proportion of respiring (primarily fruit) to
photo synthesizing (leaf) tissue. Across periods, LAR was consistently greater under
Hxx fertility compared to Lxx fertility; this difference appeared to account for the small
improvement in RGR observed under higher fertility (RGR is the product of NAR and
LAR). Higher LAR was due to greater leaf weight ratio (L WR), a relative measure of
leaf biomass to total plant biomass. Leaf area ratio was largely unaffected by specific
leaf area (SLA), a relative measure of leaf thickness or density; across treatments and
periods, SLA averaged approximately 260 cm2 g- l Ldw. Despite consistent numerical
trends, the effect of fertility regime on LAR was only significant for P d3. However, this
38
response appeared overstated, because LAR for the LLL and HHH border treatments
was essentially identical during the final period. Significant contrasts also appeared of
limited practical value, because the improvement in plant biomass and fruit yield was
tied to greater growth during the initial vegetative period.
Table 2 .13. Effect of fertility regime on net assimilation rate (NAR) and leaf area ratio
(LAR) of plants across each defined period of growth.
Net assimilation rate (g cm·2 week·1)
Fertility Pd1 Pd2 Pd3
regime ( 1 6-43 DAT) (44-64 OAT) (65-92 DAT)
LLL 0.004
0.001
LLH 0.008
0.002
LHL 0.005
0 .001
LHH 0.001
HLL 0.003
0 .002
HLH 0.007
0.002
HHL 0.004
0.001
HHH 0 .002
SE 0.0006 0.0004 0.0004
p -value ns ns ns
Contrasts LL vs. HH ns Lx vs. Hx ns
LLx vs. LHx ns HLx vs. HHx ns LLL vs. HHH ns Lxx vs. Hxx ns
Pd1
Leaf area ratio (cm2 g.1)
Pd2 Pd3
( 1 6-43 DAT) (44-64 DAT) (65-92 DAT)
62 46 bc
95 41 cd
59 36 e 38 de
82 53 a
1 1 7 48 ab
75 43 bcd 45 bc
8 5
ns ns
ns ns
*** ns ***
ns , *** non-significant at p < 0 . 10 , or significant at p < 0.0 1 , respectively; means within columns separated using the Least Significant Difference (LSD) test, p < 0.05
2.2 .3 Nutrient AnalYSis
2.2.3 . 1 Leaf Tissue
The youngest mature leaf (YML) was collected to monitor the nutrient status of
the aeroponic crop. This tissue corresponded to that commonly collected and analyzed
for comparison against established leaf sufficiency concentrations (Hartz et al. , 1 998).
Additionally, the concentration of N, P and K was determined on a sample of whole
plant leaves (WPL) collected from all ages and positions on the plant. The results of
WPL analyses were used in calculating nutrient uptake in the leaf (as representative of
all leaves, rather than just the newest leaves), but were otherwise not presented. As
39
expected, the concentration of N, P and K in YML and WPL tissues were consistently
correlated (p < 0.0 1 , r = 0.90, 0.47 and 0.88 , respectively).
2.2.3. J a NITROGEN CONCENTRATION
Seasonally, YML N concentration decreased under all treatment regimes (Table
2. 1 4) . In plants initially receiving low fertility, this decline was greatest as the
vegetative framework developed; in those initially receiving high fertility, decline was
greatest as fruit sinks developed. During Pdl , YML N concentration averaged 4.5 and
5 .3 % in the low and high treatments, respectively. Based on the leaf sufficiency range
suggested for early bloom, N concentration was adequate in plants receiving high
fertility, but fell increasingly below this threshold in plants receiving low fertility.
Despite consistently greater N concentration under high fertility, on only two of the four
dates during vegetative development was this effect significant (29 and 36 DAT).
Table 2.14. Effect of fertility regime on seasonal N concentration (%) in YML tissue.
N concentration (%) in YML tissue x OAT
Fertility
regime
P d1 (vegetative development) P d2 (fruit development) P d3 (ripening)
LLL LLH LHL LHH HLL HLH HHL HHH
S E p -value
22 29 36 43 50 57 64
5.7 4.5 3.9 4.0
5.8 5.2 5.1 5.2
0 . 13 0.20 0.36 0.35
ns *** ns
3.6 4.0 3.8
4.0 4 . 1 4.0
4.7 4.6 4.2
4.2 4.1 4.2
0.20 0 . 1 4 0.07 ns ns ns
80 92
3.7 3.5
3.6 3.3
3.7 4.0
3.8 3.9
3.6 3.6 4.4 3.4
4.0 3.5
4.2 4.2
0.09 0 . 1 4
ns ns
Sufficiency levels Z -------- 4.6 - 5.2 % --- -- 3.5 - 4.5 % ------ - 2. 7 - 3.B % --
Contrasts LL vs. HH Lx vs. Hx
LLx vs. LHx HLx vs. HHx LLL vs. HHH Lxx vs. Hxx
ns ns
ns ns ns ns
ns , *, **, *** non-significant at p < 0 . 1 0, or significant at p < 0 . 10 , 0 .05 or 0 .0 1 , respectively Z sufficiency levels from Hartz et al. ( 1 998)
ns ns ns
**
ns ns ns ns
During P d2, the numerical trend towards greater N concentration with high initial
fertility continued, although this difference appeared to moderate; YML N
40
concentration averaged 3 .9 and 4.3 % in the Lx and Hx treatments, respectively. On no
date was the effect of treatment significant. Contrasts on specific treatment means were
also not significant. All treatments resulted in tissue N concentrations comparable to
the sufficiency range suggested for full bloom and fruit development. During P d3 , tissue
N concentration averaged 3 .7 and 3.9 % in the Lxx and Hxx treatments, respectively;
these values continued to reflect adequate tissue N levels based on the sufficiency range
suggested for early fruit ripening. On neither sampling occasion during fruit ripening
was the effect of treatment on tissue N concentration significant. However, at 80 DAT
the contrast comparing N concentration between the Lxx and Hxx groupings was
significant; higher fertility resulted in greater tissue N. All other contrasts of interest
were not significant. Across periods, reinstating high fertility after a phase of low
fertility did not consistently increase tissue N concentration.
2.2. 3 .1b PHOSPHORUS CONCENTRATION
Seasonally, YML P concentration did not vary greatly under any treatment
regime, averaging approximately 0.77 % (Table 2. 1 5). Across all three growth periods,
P concentrations were very high compared to typical leaf sufficiency standards. On
only one date (64 DAT) was there a significant effect of treatment on tissue P; for this
sampling, the HH treatment had lower tissue P than the other treatments. However, this
difference was comparatively small and was associated with unusually low sampling
variability (cv = 3 %). Several contrasts were significant at 57 and 64 DAT; greater
tissue P was associated with lower fertility, although the difference remained small and
did not appear to be of any practical importance (values were above sufficiency
standards). All other contrasts were not significant. Across periods, reinstating high
fertility after a phase of low fertility did not consistently increase tissue P concentration.
4 1
Table 2. 15. Effect of fertility regime on seasonal P concentration (%) in YML tissue.
Fertil ity
regime
LLL LLH
LHL
LHH
HLL HLH
HHL
HHH
SE p -value
Sufficiency levels Z
Contrasts LL vs. HH Lx vs. Hx
LLx vs. LHx HLx vs. HHx LLL vs. HHH Lxx vs. Hxx
P concentration (%) in YML tissue x OAT
P d1 (vegetative development) P d2 (fruit development)
22 29 36 43 50 57 64
0.71
0.75
0.03 ns
0 .78 0.77 0.78
0.79 0.78 0.68
0.02 0.02 0.03 ns ns ns
------- 0.32 - 0.49 % -------
0.77 0.83 0.85 a
0.79 0.78 0.79 a
0.82 0.75 0.78 a
0.67 0.69 0.68
0.04 0.03 0.03 ns ns * *
--- 0.25 - 0.41 % ----
ns *** ns ns **
b
P d3 (ripening)
80 92 0.89 0 . 78 0.85 0.82 0.73 0.82 0.85 0 .83 0.81 0.77 0.84 0.76 0.65 0.73 0.80 0.79
0.03 0.03 ns ns
-- 0.23 - 0.37 % --
ns ns ns ns
ns ns ns ns
ns , * , ** , *** non-significant at p < 0 . 10, or significant at p < 0 . 10 , 0 .05 or 0.0 1 , respectively; means within columns separated using the Least Significant Difference (LSD) test, p < 0.05
Z sufficiency levels from Hartz et al. (1 998)
2.2 .3. 1c POTASSIUM CONCENTRA TION
There was no consistent seasonal trend in YML K concentration (Table 2 . 1 6).
During Pdl , tissue K levels averaged 4.4 and 5 .0 % in the low and high treatments,
respectively; these values were well above the sufficiency range reported for early
growth. Despite consistently greater K concentration under high fertility, on only one
of four dates during vegetative development was this effect significant (22 DAT).
When low fertility was supplied during fruit development, tissue K generally
decreased. This decline was not evident when high fertility was supplied during fruit
development and ripening; tissue K concentration actually increased in several of these
treatments late in the season. Across Pd2 YML K concentration averaged 4.3 and 4.8 %
in the Lx and Hx treatments, respectively; by comparison, across P d3 YML K
concentration averaged 4.3 and 5 . 1 % in the Lxx and Hxx treatments, respectively. In both these periods, all treatments had tissue values that continued to reflect adequate to
high K concentrations in the leaf. The effect of treatment was only significant at 64 and
80 DAT. On the first of these dates, RH fertility resulted in the greatest YML K
42
concentration, although this was not significantly different to LH fertility. On the
second date, most treatments resulted in higher tissue K than LLL fertility. Several
contrasts were also significant across these periods; a higher level of fertility generally
increased tissue K concentration. However, the grouped comparisons based on initial
fertility (Lx vs. lIx, Lxx vs. Hxx) were misleading due to high tissue K concentrations in
the continuous high fertility treatments.
The effect of reinstating high fertility after a period of low fertility consistently
increased YML K concentration compared to matching treatment combinations (i.e. LL
vs. LH, LLL vs. LLH, or HLL vs. HLH); this apparently reflected increased availability
of K in solution and the relatively high mobility of K into the plant. However, such
improvements did not appear related to growth and yield outcomes, because all plants
had above-normal tissue K concentrations and such increases generally came after fruit
set. Across periods, the concentration of K in YML tissue was negatively correlated
with the concentration of Mg and Ca in YML tissue (p < 0.0 1 , r = -0.5 1 and -0.45 ,
respectively); however, even the lowest concentration of Mg and Ca was within
established sufficiency norms ( 1 .0-2.2 % Mg and 1 .9-2.4 % Ca; Hartz et at. , 1 998).
Table 2 .16. Effect of fertil ity regime on seasonal K concentration (%) in YML tissue.
Fertil ity
regime
LLL LLH LHL LHH H LL H LH HHL HHH
SE p -value
Sufficiency levels Z
Contrasts LL vs. HH Lx vs. Hx
LLx vs. LHx H Lx vs. HHx LLL vs. HHH Lxx vs. Hxx
K concentration (%) in YML tissue x OAT
P d1 (vegetative development) P d2 (fruit development)
22 29 36 43 50 57 64
4.5 4 . 1 3.4 4.4 4.6 4.3 4.4
4 .6 4.4 5.0 ab
c
5.0 4.4 4.3 bc 4.9 4.9 5.2 5.2
0 . 14 0.1 5 0.33 0.35 ns ns ns
---- 2.2 - 3.5 % -------
5.0 4.9 5.4 a
0 . 1 3 0 . 14 0.30 ns ns * *
-- 2.0 - 3. 1 % --
ns ** *** ns
P d3 (ripening)
80 92 3.0 b 3.4 4.4 a 4 . 1 4 .8 a 4.0 5.3 a 5.3 4.2 ab 4.8 5.0 a 5.3 5.2 a 5.2 5.4 a 5.5
0 .23 0.27 ** ns
- 0.8 - 2.0 % -
** *
ns ns
ns
ns , *, ** , *** non-significant at p < 0. 1 0, or significant at p < 0 . 1 0, 0 .05 or 0.0 1 , respectively; means within columns separated using the Least Significant Difference (LSD) test, p < 0.05
Z sufficiency levels from Hartz et al. ( 1 998)
43
2.2.3 .2 Stem and Fruit Tissue
The concentration of N, P and K were determined for both stem and fruit
samples collected during the aeroponic experiment. The results of these analyses were
used in calculating nutrient uptake in each respective tissue. However, with the
exception of fruit concentrations at 92 DAT (Section 2.2.4.5), these results are not
otherwise presented. The concentration of N, P and K in stem or fruit tissue is rarely (if
ever) used in the context of determining a growth-limiting nutrient deficiency. This
reflects the difficulty of standardizing sampling from these tissues. Furthermore,
nutrients can accumulate in- or be redistributed from- healthy tissues as they age;
analysis of such tissue may therefore simply reflect the overall fertility history of the
plant rather than nutrient availability for new growth at the time of sampling.
2 .2.4 Fruit Yield and Qual ity Measurements
The effect of fertility regime on tomato yield and fruit quality is summarized in
Table 2. 1 7.
Table 2.17. Effect of fertility regime on tomato yield and fruit quality at 92 OAT.
Fertil ity Total Marketable BER Soluble Brix Adjusted
regime yield yield yield solids yield Brix yield Z (kg planr1 ) (kg planr1) (kg planr1) (OBrix) (kg Brix sol ids planr1 )
LLL 3.82 2 .59 0.31 4 .3 0. 1 1 0 . 15 LLH 4 .00 2.39 0.51 4 .5 0 . 1 1 0 . 16 LHL 4. 1 1 2.35 0.23 4.2 0 . 10 0 . 15 LHH 3.53 2.59 0 .21 4.4 0. 1 1 0 . 14 HLL 4.57 2.87 0.08 4.3 0 . 12 0 . 18 HLH 4.44 2 .76 0 . 13 4.6 0 . 13 0 . 19 H H L 4.39 2.99 0.21 4 .7 0 . 1 4 0 . 19 HHH 4.45 2.92 0 . 18 4 .7 0 . 14 0 . 19
SE 0 . 1 4 0 . 1 0 0.04 < 0.1 < 0.01 < 0.01 p -value ns ns ns ns ns ns
Contrasts
LLx vs. LHx ns ns ns ns ns ns H Lx vs. HHx ns ns ns ns ns ns LLL vs. HHH ns ns ns .. ns ns Lxx vs. Hxx • .. ..
Z adjusted Brix yield based on 90 % of total yield considered marketable ns , ., .. * non-significant at p < 0.1 0, or Significant at p < 0 . 1 0 or 0.05, respectively
44
2.2.4. 1 Total, Marketable and Blossom-end Rot Fruit Yield
The overall effect of fertility regime on total yield was not significant (Table
2. 1 7). This was despite a consistent numerical trend towards greater fruit yield under
high initial fertility. However, the contrast comparing total yield in the Lxx and Hxx treatments was significant; total yield averaged 3.87 and 4.46 kg planrl , respectively
(equivalent to a field-estimate of approximately 86 and 99 Mg ha- I , respectively). Yield
increase was primarily attributable to increased fruit number, as there were no
significant differences in mean fruit mass. Across treatments, mean fruit mass averaged
approximately 60 g fruirl. All other contrasts were not significant. Total fruit yield at
92 DAT was not correlated to FIN (Fig. 2 .9); this supported earlier plant biomass
observations, indicating that the timing of fertility (and in particular early fertility) was
most critical in determining overall productivity.
5.0
+-' 4.5
• c: • • co • Cl. Cl � 0 "0 4.0 0 Q)
':;' 0 :!: ::l J: co 3.5 0
-0
f-
3.0
0.6 0.8 1 .0 1 .2 1 .4 1 .6 1 .8 2.0 2.2
Fertility index
Fig. 2.9. Relationship between fertility index (FIN) and total fruit yield at 92 OAT. Treatments were Lxx (0) and Hxx (e).
Surprisingly, Tdw at the end of Pdl and Pd2 (end of vegetative and fruit
development, respectively) was not strongly correlated with total yield. This appeared
to be an anomaly (due to insufficient replication), particularly in light of the positive
correlation between LA at the end ofPdl and total yield (p < 0.02, r = 0.59); larger early
leaf areas were generally associated with higher fruit yields. Total yield was also
45
positively correlated with the average N concentration in YML tissue during P dl (P <
0.03, r = 0.57); N concentrations below 4.6 % (the lower threshold for established
sufficiency nonns during early bloom) were associated with yield outcomes below 4. 1
kg planrl (or less than approximately 90 Mg ha- I) . There was no clear correlation
between fruit yield and YML N status in subsequent periods. The correlation between
total yield and YML K concentration (p < 0.07, r = 0.47) appeared largely coincidental,
because all treatments resulted in tissue K levels above established sufficiency nonns.
As with total yield, the overall effect of fertility regime on marketable yield was
not significant (Table 2. 1 7) . However, the contrast between the Lxx and Hxx treatments
was significant; marketable yield averaged 2 .48 and 2.88 kg planrl , respectively
(equivalent to a field-estimate of approximately 55 and 64 Mg ha-I , respectively).
There was no difference in the percent of yield that was marketable at 92 DAT,
averaging approximately 65 % of total fruit yield for both the Lxx and Hxx groupings.
As expected, total and marketable yields were positively correlated (p < 0.03 , r = 0.55).
No attempt was made to correlate marketable yield with FIN or other variables of
interest because the crop was not commercially mature at 92 DA T.
The overall effect of fertility regime on the incidence of blossom-end rot (BER)
was not significant (Table 2. 1 7); BER is an imperfection that reduces marketable yields.
The contrast between the Lxx and Hxx groupings was however significant; treatments
receiving low initial fertility had approximately double the incidence of BER to those
receiving high initial fertility. Very large sampling variability was associated with BER
yield (cv = 70 %). Contrasts remained the same when BER incidence was adjusted to a
percent of total yield (to standardize for differences in productivity). There was no
correlation between FIN and percent BER yield at 92 DAT. However, percent BER was
negatively correlated with average Ca concentration in fruit tissue during P dl (P < 0.02,
r = -0.57); fruit concentrations below 0. 1 1 % Ca were associated with BER levels in
excess of 6 % of total yield. As expected, BER levels were not strongly correlated with
fruit Ca concentrations in subsequent periods (early rather than late deficiency is
thought to promote this disorder). There was no evidence to suggest that high
concentrations of N or K in YML tissue were correlated to greater BER expression.
Importantly, the incidence of BER overestimated that typical in the field, where levels
are more commonly only 1 -3 % of total yield. To some extent, cautious grading of only
slightly-affected fruit may have increased observed levels under aeroponics;
46
commercially, fruit that exhibit some darkening at the dista1 end but are not deeply
sunken may either be constituted as paste products or be used in some limited capacity.
Across treatments, combined BER and marketable yields (ripe fruit) still only
comprised approximately 72 % of total yield at 92 DA T. There was no significant
effect of fertility regime on this measure of fruit maturity. Pathological rots remained a
very small constituent of total yield, averaging less than 1 % (data not shown); the
predominant rot species were water mold (Pythium ultimum) and gray mold (Botrytis
cinerea). Remaining fruit were unripe (green). Collectively, these observations
suggested that a later harvest would have maximized the yield of ripe fruit. To estimate
when commercial maturity (95 % ripe fruit) would have been reached, projections were
made from 92 DAT using a ripening rate of 2.5 % colour change per day; this reflects
the approximate rate of ripening for tomatoes grown for processing (T.K. Hartz,
personal communication). This projected rate was also comparable to that observed
under aeroponics between 64 and 92 DAT (approximately 2 .6 % per day; data not
shown). Based on these predictions, the LLL treatment would have reached a 95 %
ripe-stage of maturity approximately 1 02 DAT; comparatively, the HHH treatment was
projected to reach this stage of maturity only 2 days later ( 1 04 DAT). Such a small
difference between the border treatments would be largely inconsequential in a
commercial setting.
2.2.4.2 Fruit Soluble Solids
Across treatments SSC of marketable fruit was low, averaging < 4.5 °Brix; an
acceptable target for fruit grown commercially in the field is > 5 .0 °Brix. The overall
effect of fertility regime on fruit SSC was not significant (Table 2 . 1 7). However,
contrasts confirmed that HHH and Hxx fertility significantly increased SSC outcomes
compared to LLL and Lxx fertility, respectively. The significant contrast between the
Lxx and Hxx regimes appeared to be overstated, because HLL fertility resulted in
identical fruit SSC to LLL fertility. An additional contrast was made specific only to
the fruit ripening period (xxL vs. xxH); high late fertility significantly (p < 0.08)
increased fruit SSC levels. It appeared that these effects were in part additive (i.e. high
fertility during ripening increased SSC, but to a greater level when high fertility was
also supplied during fruit development).
47
Fruit SSC was not correlated to total yield, but was to FIN (p < 0.0 1 , r = 0.65; Fig
2. 1 0); a higher index generally resulted in greater fruit Brix. Leaf area (as an indicator
of assimilate supply in plants) at the end of Pdl was positively correlated with fruit SSC
at 92 DAT (p < 0.05, r = 0.5 1 ) ; LA at the end of P d2 and P d3 was an increasingly poor
predictor of SSC potential. Fruit SSC was positively correlated with fruit K concentration at 92 DAT (p < 0.05, r = 0.52); however, the overall improvement in SSC
was comparatively small (for each 0.5 % increase in fruit K concentration, fruit SSC
increased by < 0. 1 °Brix).
4.8
4.7 - • " , 'et
4.6 - . .. -/' r = 0 .65 ..-.. >< ·c 4.5 -aJ o
0 ......... 0 4.4 Cl) o Cl) .-':; 4.3 - 0 '- • u..
4.2 o
4. 1
4.0 0 .8 1 .0 1 .2 1 .4 1 .6 1 .8 2.0
Fertility index
Fig. 2 .10. Relationship between fertility index (FIN) and fruit soluble solids concentration (SSe) at 92 OAT. Treatments were Lxx (0) and Hxx (e).
2.2.4.3 Brix Yield and Adjusted Brix Yield
The overall effect of fertility regime on Brix yield (marketable yield x fruit SSC)
was not significant (Table 2. 1 7). However, the contrast comparing Brix yield in the Lxx
and Hxx treatments was significant; Brix yield was 0. 1 1 and 0. 1 3 kg Brix solids planrl ,
respectively (equivalent to a field-estimate of approximately 2.4 and 2 .9 Mg Brix solids
ha- I , respectively) . Brix yield was more closely correlated with marketable yield (p <
0.0 1 , r = 0.95) than fruit SSC (p < 0.07; r = 0.47).
Across treatments, Brix yields were underestimated because the final harvest at
92 DAT occurred before full maturity (i.e. > 95 % ripe fruit). An adjustment of Brix
48
yield was subsequently made to reflect a typical field maturity; marketable yield was
considered as 90 % of the total yield measured at 92 DAT (assuming 5 % greens and 5
% culls, normal field levels). Although consistently higher, adjusted Brix yield
remained unaffected by fertility regime (Table 2 . 1 7) . Contrast trends were identical to
those for the unadjusted data. Adjusted Brix yield for the Lxx and Hxx treatments was
0. l 5 and 0. 1 8 kg Brix solids planrl , respectively (equivalent to a field-estimate of
approximately 3.7 and 4. l Mg Brix solids ha- I , respectively). Original and adjusted
Brix yields were both positively correlated with FIN (P < 0.05 and 0.08, r = 0 . 5 1 and
0.49, respectively; Fig. 2. 1 1 a, b).
0.20 (a) (b)
• • !' 0.18 '- • r = 0.49 c
et! 0.. en 0.16 - . · · ·6
"0 .. · ·/ 0 0 en o· x
·c 0. 1 4 - • 0 cc •• 0> . .::tt:. . . .•
. . � = 0.51 "0 0.12 • Cl> 0 >= 0 x 0
·c 0.10 cc 0
0.08 I I I I I 0.8 1 .0 1 .2 1 .4 1 .6 1 .8 2.0 0.8 1 .0 1 .2 1 .4 1 .6 1 .8 2 .0
Fertility index
Fig. 2 .1 1 . Relationship between fertility index (FIN) and actual Brix yield (a) and adjusted Brix yield (b) at 92 OAT. Treatments were Lxx (0) and Hxx (e). Adjusted Brix yields were calculated using a revised marketable yield estimate (90 % of total yield at 92 OAT).
2.2.4.4 Titratable Acidity and Lycopene Content
Titratable acidity (TA) was not significantly affected by fertility regime, varying
over a comparatively narrow range (Table 2. 1 8). Across treatments, TA averaged 0.43
%; high acidity combined with high sugars is generally associated with ideal fruit
flavour. The contrast comparing TA between the two border treatments was not
significant, despite a small numerical increase (approximately 0.04 %) under continuous
49
high fertility. The contrast between Lxx and H.xx fertility was not considered relevant
because it averaged fruit quality response during P d2 and P d3 (these are the periods in
which fruit development and ripening occur, and therefore when fertility regime would
most likely have affected TA). Two additional contrasts were made to specifically
detail the effect of late season fertility (xxL vs. xxH, and xLL vs. xHH). Despite
numerically greater TA with higher late-season fertility, neither contrast was significant.
However, FfNAdj. was positively correlated to TA at 92 DAT (p < 0.05 , r = 0.53). Fruit
K concentration at 92 DAT was not significantly correlated to TA; other minerals are
not thought to be directly associated with fruit acidity.
Table 2.18. Effect of fertility regime on titratable acidity (TA) and Iycopene content of fruit
at 92 OAT.
Fertility regime Titratable acidity Z Lycopene content
(%) (mg kg-1 fresh weight)
LLL 0.40 1 1 4 LLH 0.43 1 09
LHL 0.44 1 00
LHH 0.46 1 03
HLL 0.40 99
HLH 0.43 1 0 1 HHL 0.41 1 1 4
HHH 0.44 1 1 3
SE < 0.01 4 p -value ns ns
Contrasts LLL vs. HHH ns ns xx L vs. xx H ns ns x LL vs. x HH ns ns
Z expressed as the equivalent mass of citric acid as a percent of total dry mass; citric acid is the predominant organic acid in ripe tomatoes
ns non-significant at p < 0. 1 0
The effect of fertility regime on lycopene content was also not significant (Table
2. 1 8). Across treatments, lycopene content averaged approximately 1 07 mg kg-l
fresh
weight. All contrasts of interest were not significant. Lycopene content was not
significantly correlated to FfNAdj, nor was there any correlation to fruit K concentration
at 92 DAT; other fruit minerals are not thought to be directly associated with fruit
carotenoids.
50
2.2.4.5 Fruit Nutrient Concentrations
The overall effect of fertility regime on the concentration ofN, P, K, Ca and Mg
III fruit at 92 DAT was not significant (Table 2 . 1 9) ; across treatments, fruit
concentrations averaged 3 .0 %, 0.65 %, 4.9 %, 0. 1 6 % and 0.29 % for each nutrient,
respectively. Several contrasts were significant; with the exception of that for fruit K concentration (HHH resulted in greater fruit K than LLL fertility), the remaining
contrasts reflected comparatively small differences in nutrient concentration. Fruit
tissue concentrations at 92 DAT were not significantly correlated to total yield.
Table 2 .19. Effect of fertility regime on N, P, K, Ca and Mg concentration (%) in fruit tissue
at 92 OAT.
Fertil ity regime Fruit tissue concentrations (%)
N P K Ca Mg
LLL 2.9 0.62 4.0 0 . 1 9 0 .25 LLH 2.8 0.59 4.4 0.1 8 0 .25 LHL 2.9 0.71 4.8 0 . 1 8 0.27 LHH 2.8 0.63 5.5 0 . 1 5 0.27 HLL 2.8 0.60 4.6 0 . 1 5 0 .28
HLH 3 . 1 0.66 5.0 0.1 5 0.34 HHL 3.2 0.69 5.3 0 . 1 3 0 . 33 HHH 3.2 0.68 5.9 0 . 1 5 0 .34
SE 0.07 0 .01 0 . 1 8 < 0 .01 0 .01 p -value ns ns ns ns ns
Contrasts LLL vs. HHH ns ns ns
Lxx vs. Hxx ns ns ns ** ns , * , ** non-significant at p < 0 . 10 , or significant at p < 0 . 1 0 or 0 .05, respectively
2.2.5 Nutrient Uptake
Nutrient uptake in leaf, stem and fruit tissue was calculated using the biomass
and N, P and K concentration of each respective constituent; total uptake was then
combined uptake into these three plant tissues. Root uptake could not be calculated
because of previously mentioned difficulties in determining nutrient concentrations in
this tissue. However, root uptake (as distinct from the process of absorption) is
generally considered to constitute only a very small percentage of total nutrient
accumulation in processing tomato (Ward, 1 964). This primarily reflects the small
contribution of roots to overall plant biomass at harvest.
5 1
Cumulative uptake of N, P and K was calculated at the end of each defined
growth period. This approach provided a convenient basis on which results could be
applied in the field, where from a practical stand-point growers would have some
consistency in fertilizer applications (rather than a new regime every week). Because
plant growth and subsequent fruit yield was largely determined according to fertility
regime during the initial period, nutrient uptake appeared best matched to the grouped
Lxx and Hxx treatments; given the limited number of replications per treatment, this
also improved the overall accuracy of uptake estimates for maximum productivity.
2.2.5 . 1 Nitrogen Uptake
Seasonally, total N uptake increased under all treatments (Table 2 .20). Nitrogen
uptake was strongly correlated with total plant biomass (p < 0.0 1 , r = 0.98); greater total
biomass resulted in higher cumulative N uptake. Across periods, there was a consistent
numerical trend towards greater total N uptake resulting from high initial fertility.
However, the overall effect of treatment was only significant at 92 DAT. On this date,
cumulative N uptake was greatest in plants that had received HHH fertility, and lowest
in those that had received LLL, LHL and LHH fertility. Contrasts simplified the
general trends observed at 92 DAT; both HHH and Hxx fertility regimes resulted in
significantly greater N uptake than LLL and Lxx fertility, respectively. Seasonally, total
N uptake was approximately 8 .6 and 1 2. 1 g planf' for the grouped Lxx and Hxx
treatments, respectively; this was equivalent to field uptake in the above-ground
constituents of approximately 1 9 1 and 269 kg N ha- ' , respectively. For maximum
productivity (i.e. Hxx fertility), the plant required approximately 2 .7 kg N per
megagram of fresh fruit yield.
Greater total uptake was the result of increased N accumulation in both
vegetative (leaf and stem) and fruit tissues. However, across periods the percent of total
plant N in either vegetative or fruit tissue was not significantly different between the
Lxx and Hxx fertility regimes. Across treatments, vegetative uptake averaged 94, 60
and 4 1 % of total N accumulated at 43, 64 and 92 DAT, respectively; fruit uptake
therefore accounted for 6, 40 and 59 % of total N accumulated, also respectively.
Individually, there remained a strong correlation between leaf, stem and fruit
biomass and N uptake into each separate constituent (p < 0.0 1 , r = 0.96, 0.98 and 0.98,
respectively). Uptake of N into leaf and stem tissues largely stopped when vegetative
52
Table 2.20. Effect of fertil ity regime on cumulative N uptake (total, leaf, stem and fruit) at the end of each defined period of growth.
Fertil ity Cumulative uptake at the end of each growth period (g N planr1)
regime Total Leaf Stem Fruit
43 OAT 64 OAT 92 OAT 43 OAT 64 OAT 92 OAT 43 OAT 64 OAT 92 OAT 43 OAT 64 OAT 92 OAT
LLL 6.5
8.6 bc 2.3
2.3 1 . 1
1 .0 c 3.1
5.3
LLH 3.5
9.6 abc 2.5
2.5 0.8
1 .6 bc 0.2
5.5
LHL 7.9
8.0 c 2.9
1 .4 1 .7
1 . 1 c 3.3
5.5 LHH 8.2 bc 1 .7 1 .6 bc 4.9
HLL 9.0
1 1 .8 abc 3.8
3.2 2.3
2.5 ab 2.9
6.1
HLH 4.7
1 2.0 ab 3.2
2.3 1 .2
3.3 a 0.3
6.4
HHL 9.7
1 1 .3 abc 3.8
1 .9 2 .3
2.6 ab 3.6 6.8
HHH 1 3. 1 a 3.3 2.3 ab 7.5
SE 0.48 0.65 0.56 0.32 0.36 0.27 0 . 16 0.24 0.22 0.05 0.20 0.28 p -value ns ns * ns ns ns ns ns ns ns ns
Contrasts VI LL vs. HH ns ns ns ns w Lx vs. Hx ns ns ns ns
LLx vs. LHx ns ns ns ns HLx vs. HHx ns ns ns ns LLL vs. HHH ** ns ** Lxx vs. Hxx *** ns ns , *, **, *** non-significant at p < 0.1 0, or significant at p < 0 . 10 , 0.05 or 0.01 , respectively; means within columns separated using the Least Significant Difference (LSD) test, p < 0.05
growth truncated mid-season. Across periods, the effect of fertility regime on leaf
uptake was not significant. This was despite a trend towards greater leaf N uptake at
both 43 and 64 DAT in plants that had initially received high fertility. Differences in
leaf uptake at 92 DAT were confounded by variability in leaf biomass and also net
redistribution of N to the fruit. The effect of fertility on stem N uptake was only
significant at 92 DAT. Contrasts clarified the general response to fertility at this time;
greater stem N uptake was found with HHH and Hxx fertility compared to LLL and Lxx
fertility, respectively. Stem N uptake in the HLH treatment appeared
uncharacteristically high at 92 DAT, due primarily to high stem biomass. On no date
was the overall effect of fertility on fruit N uptake significant. However, contrasts at 92
DAT also confirmed the trend towards higher N uptake with HHH and Hx.x fertility
compared to LLL and Lxx fertility, respectively.
2.2.5.2 Phosphorus Uptake
Seasonally, total P uptake increased under all treatments (Table 2.2 1 ).
Phosphorus uptake was strongly correlated with total plant biomass (p < 0.0 1 , r = 0.98);
greater total biomass resulted in higher cumulative P uptake. Differences in total P
uptake were comparatively small at 43 and 64 DAT. However, the overall effect of
treatment on P uptake was significant at 92 DA T. On this date, cumulative P uptake
was greatest in plants that had received HLH, HHH and HLL fertility, and lowest in
those that had received LLL and LHL fertility. Contrasts simplified the general trends
observed at 92 DAT; both HHH and Hxx fertility regimes resulted in significantly
greater P uptake than LLL and Lxx fertility, respectively. Seasonally, total P uptake
was approximately 2 . 1 and 3 .0 g planrl for the grouped Lxx and Hxx treatments,
respectively; this was equivalent to field uptake in the above-ground constituents of
approximately 47 and 67 kg P ha- I , respectively. For maximum productivity, the plant
required approximately 0.67 kg P per megagram of fresh fruit yield.
Greater total uptake was the result of increased P accumulation In both
vegetative and fruit tissues. However, across periods the percent of total plant P in
either vegetative or fruit tissue was not significantly different between the Lxx and Hxx
fertility regimes. Across treatments, vegetative uptake averaged 89, 66 and 48 % of
total P accumulated at 43, 64 and 92 DAT, respectively; fruit uptake accounted for 1 1 ,
34 and 52 % oftotal P accumulated, also respectively.
54
VI VI
Table 2.21 . Effect of fertility regime on cumulative P uptake (total, leaf, stem and fruit) at the end of each defined period of growth.
Fertil ity regime
43 OAT
LLL LLH
0.9 LHL LHH HLL HLH
1 .0 HHL HHH
SE 0.06 p -value ns
Contrasts
LL vs. HH Lx vs. Hx
LLx vs. LHx HLx vs. HHx LLL vs. HHH Lxx vs. Hxx
Total
64 OAT
1 .8
2.0
2 . 1
2.2
0 . 14 ns
ns
ns
92 OAT
2.0 c 2.4 abc 2.0 c 2 . 1 bc 3.0 ab 3.1 a 2.7 abc 3.0 a
0 . 1 3
ns
ns ** ***
Cumulative uptake at the end of each growth period (g P planf1)
Leaf Stem Fruit
43 OAT 64 OAT 92 OAT 43 OAT 64 OAT 92 OAT 43 OAT 64 OAT
0.7 0.5
0.4 0.4 c
0.7
0.5 0.6
0.3 0.6 bc
0.1
0.7 0.3
0.6 0.4 c
0.7 0.4 0.6 bc
0.8 0.8
0.7 0.9 ab
0 .6
0.6 0.6
0.3 1 . 1 a
0.1
0.8 0.4
0.7 0.8 ab
0.7 0.7 0.7 bc
0.03 0 .08 0.06 0.03 0.06 0.06 0.01 0.04
ns ns ns ns ns ns ns
ns ns ns
ns ns ns
ns ns
ns * ns * ns
92 OAT
1 . 1
1 .2
1 .3 1 . 1
1 .3
1 .4
1 .5 1 .6
0.05 ns
ns
ns *
**
ns , *, **, *** non-significant at p < 0 . 1 0, or significant at p < 0 . 10 , 0.05 or 0.01 , respectively; means within columns separated using the Least Significant Difference (LSD) test, p < 0.05
Individually, there remained a strong correlation between leaf, stem and fruit
biomass and P uptake into each separate constituent (p < 0.0 1 , r = 0.96, 0.98 and 0.99,
respectively. As with N, uptake of P into leaf and stem tissues largely stopped when
vegetative growth truncated mid-season. Despite a small numerical trend towards
greater leaf P uptake in plants that had initially received high fertility, on no date was
this effect significant. The effect of fertility on stem P uptake was only significant at 92
DAT. Contrasts clarified the general response to fertility at this time; greater stem P
uptake was found with HHH and Hxx fertility compared to LLL and Lxx fertility,
respectively. The significant contrast between HLx and HHx fertility appeared to be of
little practical relevance, and was influenced by the higher stem biomass (and therefore
uptake) observed for the HLH treatment. On no date was the overall effect of fertility
on fruit P uptake significant. However, contrasts at 92 DAT confirmed the trend
towards higher P uptake with HHH and Hxx fertility compared to LLL and Lxx fertility,
respectively.
2 .2 .5 .3 Potassium Uptake
Seasonally, total K uptake increased under all treatments (Table 2 .22).
Potassium uptake was strongly correlated with total plant biomass (p < 0.0 1 , r = 0 .96);
greater total biomass resulted in higher cumulative K uptake. Across periods, there was
a consistent numerical trend towards greater total K uptake resulting from high initial
fertility. However, the overall effect of treatment was only significant at 92 DAT. On
this date, cumulative K uptake was greatest in plants that had received HHH , HHL,
HLH and HLL fertility and lowest in those that had received LLL and LHL fertility.
Contrasts simplified the general trends observed at 92 DAT; both HHH and Hxx fertility
regimes resulted in significantly greater K uptake than LLL and Lxx fertility,
respectively. Seasonally, total K uptake was approximately 1 3 .8 and 20.8 g planrl for
the grouped Lxx and Hxx treatments, respectively; this was equivalent to field uptake in
the above-ground constituents of approximately 307 and 462 kg K ha-I , respectively.
For maximum productivity, the plant required approximately 4.7 kg K per megagram of
fresh fruit yield. The ratio of total K : N uptake in the plant was approximately 1 .7 : I
under Hxx fertility.
Greater total uptake was the result of increased K accumulation in both
vegetative and fruit tissues. However, across periods the percent of total plant K in
56
Table 2.22. Effect of fertil ity regime on cumulative K uptake (total, leaf, stem and fruit) at the end of each defined period of growth.
Fertil ity Cumulative uptake at the end of each growth period (g K planf1)
regime Total Leaf Stem Fruit
43 OAT 64 OAT 92 OAT 43 OAT 64 OAT 92 OAT 43 OAT 64 OAT 92 OAT 43 OAT 64 OAT 92 OAT
LLL 9.0
12 .4 c 2.6
2.7 1 .B b 2.3 c
4.7 7 .4 b
LLH 4.4
1 4.9 bc 2.7
3.1 1 .4
3.3 bc 0.3
B.5 b LHL
1 1 .8 1 2.9 c
4.0 1 .8
2.5 ab 2.3 c 5.3
B.B b LHH 1 5.1 bc 2.6 3.1 bc 9.4 b HLL
1 3.0 1 9.4 ab
5.B 5.0
3.1 ab 4.4 ab 4.1
1 0.0 b HLH
5.9 20.0 ab
3.6 3.7
2.0 5.9 a
0.4 1 0.4 ab
HHL 1 4.6
1 9.3 ab 5.7
3.0 3.9 a 5.1 a
5.0 1 1 .2 ab
HHH 24.4 ab 5.3 5. 1 a 1 4.0 a
SE 0.54 1 .0B 1 .0B 0.30 0.63 0.41 0.21 0 .34 0.35 0.05 0.39 0.57 p -value ns ns ns ns ns ns ns ns
Contrasts VI LL vs. HH ns ns ns -....l Lx vs. Hx ns ns ** ns
LLx vs. LHx ns ns ns ns HLx vs. HHx ns ns ns * LLL vs. HHH *** ns *** *** Lxx vs. Hxx *
ns , *, **, *** non-significant at p < 0. 10 , or Significant at p < 0 . 1 0, 0.05 or 0.01 , respectively; means within columns separated using the Least Significant Difference (LSD) test, p < 0.05
either vegetative or fruit tissue was not significantly different between the Lxx and Hxx
fertility regimes. Across treatments, vegetative uptake averaged 93, 59 and 42 % of
total K accumulated at 43, 64 and 92 DAT, respectively; fruit uptake accounted for 7,
41 and 58 % of total K accumulated, also respectively.
Individually, there remained a strong correlation between leaf, stem and fruit
biomass and K uptake into each separate constituent (p < 0.0 1 , r = 0.96, 0.9 1 and 0.97,
respectively). As with N and P, uptake of K into leaf and stem tissues largely stopped
when vegetative growth truncated mid-season. Across periods, the effect of fertility
regime on leaf uptake was not significant. This was despite a trend towards greater leaf
K uptake at both 43 and 64 DAT in plants that had initially received high fertility.
Differences in leaf uptake at 92 DAT were confounded by variability in leaf biomass
and also uptake of K not apparently associated with a critical growth requirement.
However, at 92 DAT the contrast between the grouped Lxx and Hxx treatments was
significant, with greater overall leaf K uptake with higher fertility. The effect of
fertility on stem K uptake was significant at both 64 and 92 DAT. On both dates,
contrasts simplified the general response to fertility; HHH and Hxx treatments resulted
in greater stem K uptake than LLL and Lxx treatments, respectively. As with both N
and P, high stem biomass at 92 DAT inflated stem K uptake for the HLH treatment.
The effect of fertility regime on fruit uptake was only significant at 92 DA T. Contrasts
confirmed that HHH and Hxx fertility resulted in greater fruit K uptake than LLL and
Lxx fertility, respectively. The significant contrast between HLx and HHx fertility
appeared to be of little practical relevance, driven primarily by very high K uptake
under the HHH treatment.
2.3 Discussion
2.3 . 1 Effect of Fertil ity on Plant Biomass and Growth
High fertility during the initial period of growth increased the vegetative
framework established. This difference in early vegetative growth largely determined
subsequent yield outcomes. Fruit biomass was maximized with high initial fertility,
primarily due to greater fruit number. Improved fruit number likely resulted from
greater assimilate supply in plants that established larger leaf canopies; the main source
of assimilates for fruit growth come from the leaf (Ho and Hewitt, 1 986). This
58
appeared supported by the positive correlation between fruit yield and leaf area at the
end of the first period of growth. These observations under aeroponic conditions were
consistent to those made in the field; Colla et al. ( 1 999) found that higher fertility
resulted in a larger leaf canopy, which increased flowering, fruit number and subsequent
fruit yield.
The relationship between fruit yield and assimilate capacity reflected the growth
habit in processing tomato. Determinate cultivars have been bred to flower and set fruit
over a short period of time (Atherton and Harris, 1 986). Whilst this maximizes the
yield of ripe fruit during once-over destructive harvesting, plants rarely have excess
assimilate supply (Stevens, 1 976) . Because assimilates are preferentially directed to
developing fruit (Ho, 1 984; Wardlaw, 1 990), when demand equals or exceeds supply
not only does new vegetative growth stop, but later-setting flowers may abort (reducing
fruit number). This phenomenon appears to be reflected in the fact that the majority of
fruit yield is carried on earlier rather than later flower trusses (Julian, 1 990; Renquist
and Reid, 1 998). Not surprisingly then, reinstating adequate fertility after vegetative
growth had stopped and fruit number had been determined did not improve yield. A
complete discussion of fruit yield parameters is provided in Section 2.3 .3 .
Greater accumulation of vegetative biomass with high initial fertility was
attributed to an increased RGR in plants. Although this difference was comparatively
small, it nonetheless was consistent with the overall increase in plant biomass. Growth
rates of plants during this initial phase of vegetative development were lower than
reported previously under N.Z. conditions; Pan et al. ( 1 999) found that during the first 8
weeks after field transplanting RGR averaged approximately 1 .4 g g-I week- I ; under
aeroponics, this rate was closer to 0.8 g g- I week- I . This in part reflected the omission
of biomass gain during the first 1 5 DAT in aeroponics (the initial treatment period was
between 1 6-43 DAT). The relative gain per unit of established plant biomass is greatest
early in the season and declines thereafter. The difference in RGR may have also
reflected solar radiation levels under glass, which were most likely lower than in the
field; plant growth rates are strongly linked to radiation interception (Bruggink and
Heuvelink, 1 987). In a well maintained modern glasshouse, transmittance may only be
reduced by 1 0- 1 5 %, while in older glasshouses with greater structural interferences this
reduction may be closer to 20-30 % (T.K. Hartz, personal communication). No
measurements of solar radiation levels inside the glasshouse were made during this
experiment.
59
Higher RGR was itself related to a higher LAR early in the season. These
observations were consistent with the larger leaf area that developed with high initial
fertility. Greater leaf area is thought to improve radiation interception and therefore
photosynthetic activity (Greenwood, 2001 ; Sinclair and Horie, 1 989; Tei et al. , 2002); it
is through the process of photosynthesis that new assimilates are made for continued
growth and development. These findings under aeroponics were similar to those
reported in the field by Cavero et al. ( 1 999) and Pan et al. ( 1 999); both found that
increased growth was primarily the result of a higher LAR.
Early-season LAR dynamics appeared strongly related to the N status of plants;
low initial fertility resulted in a mild N deficiency when compared to established
sufficiency norms (Hartz et al. , 1 998). This deficiency was apparently growth-limiting
to plants. A number of other studies have also documented the importance of adequate
N availability (Adams, 1 986, Cavero et al. , 1 997, 1 998; Cerne, 1 990; Clark et al. , 1 999;
Colla et al. , 1 999, 200 1 ; Greenwood et al. , 1 99 1 ) ; all reported that where N was
limiting, plant growth declined resulting in lower fruit yields. The P and K status of
plants appeared unrelated to differences in productivity under aeroponics. A complete
discussion of Ieaf nutrient concentrations is provided Section 2.3.2.
2.3.2 Effect of Ferti l ity on Leaf Nutrient Concentrations
2.3.2. 1 Nitrogen Concentration
Seasonally, leaf N concentrations declined as assimilates were redistributed to
developing fruit. This observation was consistent with previous studies conducted in
the field (Colla et al. , 2001 ; Hartz et al. , 1 998; Hochmuth et al. , 1 99 1 ; Lorenz and
Tyler, 1 983; Tei et aI. , 2002). Provided fruit demand does not result in extensive
mining of nutrients from leaves early in the season (when vegetative growth is still
important), management to allow for such a seasonal decline may reduce the need for
mid- and late-season N applications. However, the ability to accurately predict such
dynamics appears difficult; seasonal sufficiency norms have been developed as
' indicators' of overall crop nutrient status, although by the time a deficiency is detected,
new growth may already have been limited.
In plants receiving low fertility during vegetative development, the
concentration of N in leaf tissue fell increasingly below such sufficiency norms. Hartz
60
et al. ( 1 998) suggested that maximum fruit yields are typically associated with leaf N
concentrations during early bloom between 4.6-5.2 %; under aeroponics, tissue values
were as Iow as 3 .9 % with Iow initial fertility, appearing to be at least mildly deficient.
These observations were consistent with reduced vegetative growth; N is a key
constituent in chlorophyll and amino-acids, and therefore plays a critical role in leaf
development (Mengel and Kirkby, 1 987). Early N deficiency appeared to primarily
reflect high demand for this nutrient during the intensive period of vegetative growth;
the low fertility regime was unable to meet this demand. During fruit development and
ripening, tissue N concentrations were above established sufficiency norms for all
treatments. However, fruit yield dynamics had already been largely determined by this
time.
2 .3 .2 .2 Phosphorus Concentration
Seasonally, the concentration of P in leaf tissue remained very high for all
treatments; values were well above established sufficiency norms recommended by
Hartz et al. ( 1 998) for maximum yield. Leaf P did not decline during fruit
development, as has been commonly reported in the field (Hartz et al. , 1 998; Hochmuth
et al. , 1 999; Lorenz and Tyler, 1 983). This observation suggested that P availability in
solution was sufficient for optimal growth, even under low fertility; irrespective of the
concentration in solution, P was continuously available in highly-soluble plant forms.
This is in contrast to many field settings, where soil dynamics largely control P
availability; factors such as soil temperature, soil moisture, pH, salinity, compaction and
aeration can all influence equilibrium reactions and therefore plant-available P (Brady,
1 990).
This 'ready' availability throughout the season may also account for the high
concentrations found under aeroponics. Typical leaf P concentrations in processing
tomatoes range from 0.2-0.5 % (Hartz et al. , 1 998), while under aeroponics
concentrations consistently exceeded 0.7 %. Although Hochmuth et al. ( 1 999) found
early-season tissue P concentrations as high as 0.9 %, this appeared to not only reflect
the very high fertilization rates used (up to 200 kg P ha-I), but also the type of tomato
grown (fresh market rather than processing); the two have dissimilar growth habits, and
can therefore have differing nutrient requirements. There was no evidence to suggest
that such high concentrations in plants grown under aeroponics reflected an actual
6 1
critical growth requirement. Differences m plant growth and fruit yield therefore
appeared unrelated to P status in the plant.
2.3 .2.3 Potassium Concentration
Seasonally, the concentration of K in leaf tissue remained very high for all
treatments; values were well above established sufficiency norms recommended by
Hartz et al. ( 1 998). There was no evidence to suggest that differences in vegetative
biomass and subsequent fruit yields were related to tissue K status. This indicated that
K availability in solution was sufficient for optimal growth, even under low fertility.
For most treatments, leaf K concentration did not decline as fruit developed.
This was in contrast to findings reported in the field, where remobilization of K to the
developing fruit causes leaf concentrations to decrease (Besford and Maw, 1 975; Hartz
et al., 1 998, 1 999, 2005; Pan et al. , 1 999; Widders and Lorenz, 1 982). In some
treatments (particularly those receiving high late fertility), leaf K concentrations
actually increased as fruit developed and ripened. This increase likely reflected luxury
absorption of K due to its higher availability in solution rather than an actual plant
demand; fruit yields were determined before such increases. A number of crops
including tomato are known to continue to absorb K well in excess of critical
requirements (Bartholomew and Janssen, 1 929; Brady, 1 990; Fisher et al., 2002; Pan et
al., 1 999; Pujos and Morard, 1 997); while it does not typically result in K toxicity, it
may overestimate plant requirements for this nutrient.
2.3.3 Effect of Fertil ity on Fruit Yield and Qual ity
2 .3 .3 . 1 Total, Marketable and Blossom-end Rot Fruit Yield
Fruit yields were greatest with high initial fertility. Higher yield was due to
greater fruit number rather than mean individual fruit mass. Increased fruit number was
previously attributed to greater assimilate capacity in plants receiving high initial
fertility, although no measurements were made to distinguish whether this effect was
related to increased flower number and/or an improvement in the percent of flowers that
successfully set. Colla et al. ( 1 999) suggested the former, reporting that flower number
per unit area increased with higher fertility; subsequent fruit number per unit area
62
increased by a similar amount. hnportantly, fruit rots (excluding BER) remained a very
small constituent of yield « 1 % across treatments). This confirmed that differences in
fruit number were not likely due to fruit that had disintegrated and were not otherwise
counted (if this were problematic, rots in general would have been higher). Marketable
yields were low for all treatments because the final harvest occurred before full
maturity. However, because high initial fertility did not increase fruit culls nor did it
cause a significant delay in crop maturity, maximizing total yield would therefore also
have maximized marketable yield.
Blossom-end rot incidence was high across treatments. This apparently
reflected the glasshouse production environment. Higher temperatures may have
promoted rapid cell expansion and fruit growth, while elevated humidity likely reduced
transpiration and water movement (with dissolved Ca) into developing fruit; both of
these events have been previously associated with localized Ca deficiency at the distal
end of fruit, which has resulted in greater BER (Adams and Ho, 1 992; Paiva et al. ,
1 998) . Blossom-end rot levels in the field are typically much lower, and more
commonly aggravated by intermittent water stress rather than low soil Ca availability;
such water stress did not appear related to BER incidence under aeroponics, because
plant roots received continual misting.
Importantly, there was no evidence to suggest that high fertility caused greater
BER. The incidence of this disorder was in fact further aggravated under low fertility,
particularly during the initial period. Tissue Ca concentrations in early developing fruit
grown under low fertility were similar to those suggested by Grierson and Kader ( 1 986)
to cause an increase in BER symptoms « 0.08 % Ca). This apparently reflected low Ca
availability in solution. Massey et al. ( 1 984) suggested that BER incidence increased
with solution concentrations below 70 mg L-1 Ca; under aeroponics, the low fertility
regime supplied only 60 mg L- 1 Ca. All nutrients ratios were otherwise considered to be
appropriately balanced for successful greenhouse production (Tregidga et al. , 1 986).
2 .3 .3 .2 Fruit Soluble Solids
Across treatments, fruit SSC was low; high SSC (> 5.0 °Brix) is desirable, as it
improves paste yield and overall processing efficiency (Linden, 2004). Renquist and
Reid ( 1 998) and Dadomo et al. ( 1 994b) both reported high SSC for 'Cannery Row' at
maturity (5 .4-6.3 °Brix), indicating a good SSC potential in this variety. High yields did
63
not appear to cause low fruit SSC observed under aeroponics; this phenomenon has
been previously reported by Dumas et al. ( 1 994).
Poor SSC therefore apparently reflected an earlier than ideal harvest (before full
maturity) and also a lack of significant osmotic stress during fruit ripening. Late-season
water stresses are commonly imposed in the field to increase SSC levels (Cahn et al. ,
200 1 ; Calado et al. , 1 990; Lowengart-Aycicegi et al. , 1 999; May et al. , 1 990; May and
Gonzales, 1 999; Murray, 1 999; Renquist and Reid, 2001 ), while in glasshouse culture
high solution conductivity often achieves the same result (Cuartero and Fernandez
Munoz, 1 999; Ehret and Ho, 1 986; Ho et al. , 1 987; Sakamoto et al., 1999). Under
aeroponics, plants received continual misting of solutions directly onto the root surface,
effectively eliminating the potential for controlled water deficits. Roots had no
surrounding medium to absorb water from, so a water cutback or cutoff would have
almost certainly killed the plants. Similarly, many studies that have found large
improvements in fruit SSC by manipulating solution conductivity have done so at levels
considerably greater (> 8 dS m-I ) than the high treatment used under aeroponics (2 dS
m-I ). Whilst improving fruit quality, solution conductivities above 2 .5 dS m-I can
reduce tomato growth and therefore yield (Maas, 1 986).
It was unclear whether the small improvement m SSC with high fertility
(particularly when supplied during fruit ripening) was the result of modestly higher
conductivity (0.7 vs. 2.0 dS m-I ) or higher K availability as suggested by Lachover
( 1 972). Although higher fruit K concentration was associated with fruit SSC, this may
have simply reflected its elevated availability in solution (as noted, K is known for
luxury uptake). Also, the improvement in fruit SSC per increase in fruit K
concentration was very small ; to have any notable impact on fruit quality, the increase
in fruit K concentration required appeared not only uneconomical under field
conditions, but also impractical. Applied K can be fixed at large quantities in many
agricultural soils (Cassman et al. , 1 990).
Although K fertility is undoubtedly related to other fruit quality factors such as
internal and external blotchy ripening (Ozbun et al., 1 967; Picha and Hall, 1 9 8 1 ) and
yellow shoulder (Hartz et al., 200 1 , 2005), the effectiveness of high fertility as fruit
ripened appeared poor; plant nutrient demand after most growth stopped was low, and
fruit yield was not improved by high late-season fertility. In the field, achieving high
SSC therefore appears best addressed by late-season irrigation practices; the effect of
such techniques is summarized in Chapter 4.
64
2.3 .3 .3 Brix Yield
Brix yields were low across treatments because of poor SSC « 5 .0 °Brix) and
low marketable yields. Renquist and Reid ( 1 998) reported high yields and fruit SSC in
'Cannery Row' , with an overall productivity close to 6 Mg Brix solids ha- I ; the
commercial nonn is at least 5 Mg Brix solids ha-1 (based on a yield of 1 00 Mg ha-1 and
fruit SSC of 5 .0 °Brix). The adjustment made to marketable yields to reflect a 'typical '
commercial maturity increased Brix yields, but levels were still below 5 Mg Brix solids
ha- I ; no correction could be made for fruit SSC.
Despite being low, Brix yields were maximized with high initial fertility. This
was consistent with improved yield productivity, and to a lesser extent higher fruit SSc.
These findings were similar to those made in the field by Colla et at. ( 1 999), who
concluded that Brix yields were maximized with high N applications rates compared to
plants that received no fertilizer; this was due to both higher yield and fruit SSc. The
strategy then of maximizing early growth for high yield appeared critical in maximizing
Brix yield outcomes. Nutrient deficiencies during vegetative growth and fruit
development not only have the potential to decrease yield, but may also reduce leaf area
and therefore assimilate capacity of the plant, potentially lowering fruit SSc.
2.3 .3 .4 Fruit Titratable Acidity
High late-season fertility numerically increased titratable acidity of fruit,
although this improvement was small . Similar increases with either high fertility or
high conductivity have been reported elsewhere (Colla et al. , 1 999, 200 1 ; Cuartero and
Fermindez-Munoz, 1999; Mitchell et al. , 1 99 1 a, b). A number of researchers have
suggested that greater K availability increases fruit TA (Adams et al., 1 978; Carangal et
al. , 1 954; Davies, 1 964; Davies and Winsor, 1 967; Stevens, 1 972). The two were not
directly correlated under aeroponics, despite generally greater fruit K concentrations
with increasing fertility. However, the positive correlation between the adjusted
fertility index and TA may have reflected K availability rather than overall solution
conductivity; K availability increased at a rate directly proportional to this index (i.e. a
higher index provided high K concentration in solution for a longer duration of fruit
development and ripening).
65
Whilst there appears to be some potential to favourably influence fruit TA with
late-season fertility, it does not represent a practical field option. Under controlled
glasshouse conditions it is possible to manipulate the conductivity or fertility of a small
soil volume (i.e. pot) or solution with relative ease and affordability; in the field
however, fertilizer applications required to improve fruit quality often far exceed that
expected for agronomic yield response. A similar observation was made by Hartz et al.
(2005); although the incidence of the colour disorder yellow shoulder was reduced with
higher K fertility, without a yield increase such applications were considered
uneconomical. Because there is no price premium offered for fruit with high TA, there
is little incentive to apply fertilizer at rates above those required for maximum yield.
Furthermore, unlike fresh market tomatoes, processed tomato products can be
supplemented with citric acid if levels are too low. It would also appear that TA is
influenced by other factors at least equally, if not more so than soil fertility; these may
include fruit age, cultivar selection and the production environment (Colla et al. , 1 999,
200 1 ; Davies and Hobson, 1 98 1 ; Davies and Winsor, 1 969).
2 .3 .3 .5 Fruit Lycopene Content
There was no effect of fertility regime on the lycopene content of ripe fruit.
Levels were comparable to those reported for processing tomatoes grown in the field
(Barrett and Anthon, 200 1 ; LOpez et al. 200 1 ). Some studies have suggested a positive
link between K fertility and lycopene content (Amable and Sinnadurai, 1 977; Flores et
al. , 2004; Trudel and Ozbun, 1970, 1 97 1 ), although under aeroponics the two were not
correlated. This observation may have been influenced by high fruit K concentrations,
which averaged 4.9 % across treatments; typical fruit K concentrations are more
commonly in the range of 4.0-4.5 % (Hartz et at. , 1 999).
There was little evidence supporting the use of high late-season K fertility to
improve lycopene content. The best grower option still remains cultivar selection;
Barrett and Anthon (2001 ) reported that some commercial cultivars have almost double
the lycopene content of others. Furthermore, while fruit colour and colour uniformity
are important quality attributes for processing, they are not entirely analogous to
lycopene content; therefore, achieving acceptable and uniform colour is the pre
requisite rather than achieving high lycopene content per se. In addition to cultivar
66
selection, scheduling harvesting for when the majority of fruit are ripe (i .e. < 5 %
greens) commonly achieves acceptable fruit colour.
2 .3 .3 .6 Fruit Nutrient Concentrations
The concentration of N, P and Mg in fruit at 92 DAT appeared to be largely
inconsequential; these nutrients were not correlated to fruit quality variables of interest
(SSC, TA or lycopene content). Despite its purported association with fruit quality
(Usherwood, 1 985), K concentration in fruit was only weakly correlated to SSC, and
not directly correlated to TA or lycopene content. Higher fruit K concentrations
therefore appeared to largely reflect overall availability in treatment solutions.
Although Ca levels in early-developing fruit were correlated with BER incidence, by 92
DAT such concentration differences had disappeared; all fruit had tissue concentrations
well above the 0.08 % Ca threshold thought to aggravate the symptom (Grierson and
Kader, 1 986). This observation was not surprising, because BER typically develops on
the earliest fruit (Saure, 200 1 ). Tissue sampling at 92 DAT for mineral analysis also
excluded those fruit that were most severely affected by the disorder; additionally, BER
is a localized deficiency at the distal end of fruit (Adams and Ho, 1 993), whereas
collected tissue was cut longitudinally from the whole-fruit (effectively eliminating any
concentration gradient).
2.3.4 Effect of Fertil ity on N utrient U ptake
Current fertilizer recommendations have been largely based on determinations
made in the field, where supply of nutrients directly at the root surface can be limited by
a number of production and environmental factors. Not all applied fertilizers in such
field experiments may therefore be available to the plant, although they often become
part of the overall fertilizer requirement suggested for maximum tomato yield and/or
fruit quality. Aeroponic supply of nutrients effectively eliminated such constraints by
optimising the root environment for growth; nutrient uptake as determined using this
system may therefore represent the most appropriate measure of plant demand from
which seasonal applications can be derived. Under aeroponics, nutrient uptake
associated with the Hxx treatment (the HLL, HLH, HHL and HHH treatments)
67
represented that closest to 'optimal ' plant demand for maximum yield and to a lesser
extent fruit quality.
Seasonal plant uptake of N, P and K was sigmoidal, closely following the
pattern of biomass accumulation. Greatest nutrient uptake was therefore during the
period of rapid fruit biomass accumulation (under aeroponics, this was approximately
36 to 80 DAT). Similar observations have been reported in the field (Dumas, 1 990;
Halbrooks and Wilcox, 1 980; Wilcox, 1 993); despite differences in cultivars, growth
patterns, and growing environments, all found that nutrient uptake was greatest during
fruit growth. Matching nutrient supply more closely with plant demand is an important
step in improving fertilizer efficiency in the field (Dumas, 1 990; Hartz and Hochmuth,
1 996; Stark et al. , 1 983). Because the concentration of nutrients in surface runoff and
leachate is directly correlated with their availability in the soil (Blackmer, 1 987; Hartz
and Johnstone, 2005 ; McDowell and Sharpley, 2001 ; Sharpley, 1 995, 1 997), reducing
supply during periods when demand is low will also reduce the potential for
environmental loss.
Although adequate fertility is still required for full vegetative growth, plant
nutrient demand during this phase is invariably lower than during fruit growth. Where
practical, fertilizer application should therefore be concentrated as fruit develop.
However, in conventionally-managed fields it is difficult to side-dress fertilizers later
than approximately 60 days after emergence without causing extensive root and vine
damage and possibly reducing yields (banding fertilizers by tractor shanks requires
'open' bed-tops).
One option currently practiced is to water-run fertilizers (particularly N) with
furrow irrigation, although the efficiency of this practice is directly related to the
uniformity and efficiency of irrigation. In fields where irrigation uniformity and
efficiency is poor due to slow infiltration rates or excessive slope, this practice may not
evenly distribute nutrients. The use of drip irrigation technology offers a more effective
alternative to match supply with crop uptake, free from many of the cultural constraints
that characterize other production systems (Hartz and Hochmuth, 1 996). Drip
fertigation also enables the delivery of fertilizers directly to the active root-zone, which
is typically confined to the region wetted by the drip tape. Improving the efficiency and
timeliness of such applications may reduce the need for high fertilizer rates.
Seasonal plant nutrient uptake for high fruit yield (> 90 Mg ha- I) was equal to
1 2. 1 , 3 .0 and 20.8 g planfl for N, P and K, respectively; this was equivalent to
68
approximately 270, 64 and 460 kg ha-I , respectively. Plant N uptake was similar to
fertilizer applications made in high-yielding commercial fields; growers commonly
apply between 1 40-280 kg N ha-I (Hartz and Miyao, 1 997; Krusekopf et al. , 2002).
Although Halbrooks and Wi1cox (1 980) found lower N uptake per plant (9.4 g N plant
I ) than under aeroponics, much of this disparity appeared related to differences in
biomass productivity. In the field, Halbrooks and Wilcox ( 1 980) reported a total
biomass of 3 1 4 g planfl , while under aeroponics, the Hxx treatment resulted in a total
biomass of approximately 4 1 6 g plants-I . Adjusted for biomass, the percent of N per
unit of dry plant biomass were similar (approximately 3 % N).
Plant P and K uptake were above typical fertilizer application levels. Although
aeroponic culture provided the ideal root environment for plant growth and allowed
manipulation of fertility directly at the root surface, the interpretation of nutrient uptake
results obtained from this method appeared confounded by its own unique limitations.
All nutrients were supplied in highly absorbable fonns, so although nutrients varied in
concentration, they were readily available. In this system there was apparently luxury
uptake of both P and K. Eymar et al. (200 1 ) concluded similarly, suggesting that in an
ideal growing environment (where water and nutrients are not limiting and roots are
healthy) nutrients can continue to accumulate in plant tissue beyond levels normally
expected or required for growth.
Even adjusted for differences III biomass productivity, P and K uptake
determined by aeroponics was still greater than that suggested by Halbrooks and Wilcox
( 1 980); in the field, they found P and K uptake was 1 .2 and 1 4.2 g planfl , respectively.
Differences in tissue concentrations reflected the greater P and K uptake per unit of
plant biomass under aeroponic conditions. Under aeroponics, the percent of P and K
per unit of dry plant biomass was approximately 0.72 % and S.O %, respectively, while
Halbrooks and Wi1cox ( 1 980) found approximately 0.38 % and 4.S %, respectively.
Although P uptake for high yield under aeroponics (67 kg P ha-I ) was not greatly
dissimilar to commercial application norms (30-60 kg P ha-I ; Hartz and Miyao, 1 997),
many growers supply P at such high rates as an 'insurance' for early growth. It is early
season P status which is often most critical to successful growth (Strand, 1 998),
particularly in fields that are planted in spring when low soil temperatures can greatly
reduce native P availability (Johnstone et al. , 200Sa).
By comparison, current grower K applications « 1 1 0 kg K ha-I ; Hartz and
Miyao, 1 997) rely on a large amount of plant K being supplied by the soil; plant K
69
demand often exceeds 300 kg K ha- 1 (Widders and Lorenz, 1 979). Although K accumulation by tomato is greater than N accumulation, the K : N ratio is typically 1 .5 :
1 (Adams, 1 986). In this study, Hxx fertility had a ratio closer to 1 .7 : 1 . Plant P and K demand for maximum fruit yield as determined by aeroponics may therefore have
overestimated that typical in the field environment. Field-testing is required to validate
the necessity of such high applications under normal production conditions.
70
CHAPTER 3: 1 999-2000 Field Nutrition Trial
A field experiment was conducted at the Fruit Crops Unit (long. 1 750 37'E; lat.
40° 23 ' S) , Massey University, Palmerston North, during the summer growing season of
1 999-2000. The aim was to evaluate under field conditions nutrient uptake results that
had been detennined previously using aeroponic culture; based on this prior study, plant
nutrient uptake for maximum fruit yield and quality was quantified. Closely matching
supply with plant demand was tested as a means of maintaining high yield in processing
tomato while improving fertilizer efficiency (applying nutrients continuously
throughout the season is thought to be more efficient than single applications). Drip
fertigation was selected as the most effective and practical means to supply nutrients to
plants in small but regular amounts. Drip also offers the ability to more closely regulate
irrigation. However, the influence of water supply on fruit yield and quality was
considered in a separate series of experiments (Chapter 4).
3.1 Materials and Methods
3 . 1 . 1 Experimental Site
The experimental site was selected in Palmerston North because of proximity to
the University; although most commercial production in N.Z. occurs in the Hawkes
Bay, the continuous fertigation required for this trial made such a location impractical.
Palmerston North has a temperate climate with wann, windy summers and cool, moist
winters (Burgess, 1 988). Daily meteorological data for pan evaporation, air and soil
temperatures and rainfall were collated from the nearest weather station (AgResearch
Grasslands Centre); this station was less than 1 km from the Fruit Crops Unit. A full
summary of climate data for the 1 999-2000 season is provided in Appendix n. Soil type was a Manawatu sandy loam (mixed, mesic, Dystric Eutrochrept) .
The soil was fonned on loamy alluvium over gravels, characterized with moderate
fertility and high penneability. The field had recently been converted from a long-tenn
apple orchard. A representative soil sample was collected before planting (0-30 cm
depth). The sample was air-dried before being ground to pass through a 2 mm screen.
Soil pH was detennined potentiometrically by pH electrode on a 1 :2 (v/v) soil:water
slurry (Blakemore et al. , 1 987). Available N was detennined colorimetrically following
7 1
extraction with 2 M potassium chloride (Hinds and Lowe, 1 980). Bicarbonate
extractable P was determined using the Olsen method for available P (Olsen and
Sommers, 1 982). Exchangeable cations (X-K, X-Ca, X-Mg and X-Na) were extracted
using 1 M neutral ammonium acetate (Thomas, 1 982); determination of X-K and X-Na
was by atomic emission and X-Ca and X-Mg by atomic absorption. Soil fertility is
summarized in Table 3. 1 .
Table 3.1 . Soil fertility status of the top 30 cm depth.
Fertility analysis
pH Olsen P (mg kg-1 ) Avai lable N (kg ha-1 ) Exchangeable K (cmol kg-1 ) Exchangeable Ca (cmol kg-1 ) Exchangeable Mg (cmol kg-1 ) Exchangeable Na (cmol kg-1 )
Z as determined by a commerical laboratory
3 . 1 .2 Production Detai ls
Level found
6.2 80
1 1 5
0.41
9.2 1 .39 0 .08
Nutrient status Z
medium high
medium low
medium medium medium
The hybrid cultivars 'Morse' and 'H225' were used in this study (seed of
'Cannery Row' was not available); both were selected as popular N.Z. standards.
'Morse' was considered to be most similar to 'Cannery Row' . Seed was supplied by
Heinz Wattie's Ltd. (Hawkes Bay, N.Z.). Seed was sown and seedlings raised in a
greenhouse. Four weeks prior to transplanting, seedlings were hardened in an outdoor
shade house. Plants were liquid fed as required using a solution containing 1 00, 34 and
1 00 mg L-! of N, P and K, respectively. Uncharacteristically high summer rainfall
during November delayed planting; standard bed preparation techniques were made
after the soil had dried out sufficiently. A rotary-hoe was used to form slightly-raised
1 .5 m wide beds, although these settled with heavy rainfall .
Transplants were hand-planted into single rows with in-row spacing of 30 cm;
this was equivalent to a density of 22,222 plants ha- ! (Fig. 3 . l a, b). Surface drip
irrigation was used to establish transplants. All plants received a small liquid feed
(equivalent to less than I kg ha- ! of N, P or K) immediately following planting, a
common commercial practice when transplanting. Heavy rain and strong winds during
the initial week after transplanting slowed early plant growth. Temperature sensors
72
were installed at 1 5 cm above- and below the soil surface in plant rows, with four
replicates per height and depth. Measurements were recorded every 1 5 minutes by
data-logger (Squirrel SQ, Grant Instruments, Dundee, UK). Heavy rainfall flooded the
data-logger mid-season and site-specific data was lost. Important production and
phenological dates are summarized in Table 3 .2 .
Table 3.2. Production and phenological dates, 1 999-2000.
Phenological stage
Sown Transplanted First flowering First fruit set First ripe fruit
October 20 December 9 January 28 February 9 March 1 7
Control for broad-leaf weed speCIes consisted of post-plant application with
Sencor® (Bayer Crop Science Ltd.) at 0.67 kg a.i ha- I . During the season regular in
row hand weeding was required. Weed control with Roundup® (Monsanto Co.) was
also used carefully between rows early in the season (due to poor control with Sencor) .
Insecticide control for tomato fruit worm (Helicoverpa armigera) consisted of a single
mid-season application with Attack® (Nufarm Ltd.) at 0.7 kg a.i ha- I . Multiple early-,
mid- and late-season fungicide applications were made alternating between Bravo®
(Syngenta Crop Protection, Inc.) at 1 . 8 kg a.i ha- I and Dithane® (DowAgro Science) at
1 . 5 kg a.i ha-I . Heavy rainfall provided ideal conditions (free moisture, cool and humid
after rain storms) for development of early blight (Alternaria solani) and late blight
(Phytophthora infestans). High moisture availability also resulted in a large incidence
of pathological rots as fruit ripened.
3 . 1 .3 Experimental Design
Experimental design was a split-plot within a randomized complete block, with
the main plot representing fertilizer treatment, the split-plot representing cultivar. The
two cultivars used ('Morse' and 'H225') were chosen to represent early and mid crop
maturities, respectively. Individual split-plots were one bed wide and 1 9. 8 m long,
73
Fig. 3.1a. Early-season vegetative growth at 26 OAT.
Fig. 3.1 b. Mid-season vegetative growth at 55 OAT.
74
and were replicated three times. Fertilizer treatments were initiated 1 0 DAT,
corresponding to early vegetative growth. Phosphorus was incorporated preplant (the
commercial norm); only a small quantity ( 1 0 kg P ha- I ) was applied to all treatments
equally because of the very high Olsen-P test result, which was well above the typical
agronomic threshold for processing tomato ( 1 2- 1 5 mg kg-1 P ; Hartz and Miyao, 1 997).
Seasonal application rates of N and K were based on optimum uptake that had
been determined previously under aeroponics for maximum yield and fruit quality (H.x.x fertility). Based on these results, plant demand was defined for each distinct period of
growth (vegetative development, 1 0-55 DAT; fruit development, 56-90 DAT; fruit
ripening, 9 1 - 1 20 DAT), and then divided over the duration of each period to provide a
regular application amount. Fertigation was made 3 times per week (irrespective of
rainfall), with the weekly application being split evenly across these events. Irrigation
volume was determined using a standardized evapotranspiration loss of 5 mm per day
adjusted for crop canopy development; all treatments received the same irrigation
volume. Fertilizer rates evaluated included 0 (control), 0.5, 1 .0, 1 .5 and 2.0x the
optimal plant nutrient uptake found under aeroponics. Details of N and K field
applications are summarized in Table 3.3 for the optimal treatment ( I .Ox). Fertigated N
and K (from ammonium nitrate and potassium nitrate) were injected into the irrigation
stream by water-powered proportional injectors.
Table 3.3. Optimal N and K treatments (1 .0x) applied during vegetative development, fruit
development and fruit ripening.
Optimal nutrient supply Z Plant growth stage
Vegetative development (10-55 DAT) Fruit development (56-90 DAT) Fruit ripening (91 -1 20 DAT)
Cumulative
Individual plants N K
(g) (g)
4.7 6.0 4 .7 7 .8 2 .7 7.0
1 2. 1 20.8
Field basis Y
N K
1 04 1 33 1 04 1 73
60 1 56
268 462
Z reflects optimum plant uptake for high yields (> 90 Mg ha-1 ) as determined under aeroponics (Hxx fertility); the remaining treatments were calculated as 0 (control), 0.5, 1 .5 or 2.0x these rates
Y calculated using a density of 22,222 plants ha-1
75
3 . 1 .4 Data Col lection
3 . 1 .4. 1 Leaf Analysis
Leaf samples (YML tissue) were collected from each split-plot at 55 , 74 and 99
DAT to document the N, P and K status of the crop; these dates corresponded
approximately to early bloom, full bloom and early fruit ripening ( 1 0 % of fruit
showing red colour), respectively, all of which have established tissue sufficiency
norms (Hartz et aI. , 1998). On each occasion, 1 5-20 leaves were removed per split-plot.
Leaves were rinsed to remove any surface residues, then oven dried at 65 QC for 48
hours and ground to pass through a 0 .5 mm screen. Total N, P and K concentrations
were determined as previously described.
3 . 1 .4.2 Destructive Whole-Plant Sampling
There was one destructive whole-plant harvest made 5 5 DAT, corresponding
approximately to full vegetative development (leaf and stem growth largely stops after
fruit development begins). On this date, 1 0 consecutive plants were destructively
sampled from each split-plot. Total fresh plant biomass (above-ground) of all plants
was recorded. Four plants were then randomly selected and separated into leaf and
stem tissue and each component weighed; there was no fruit biomass at 55 DAT.
Representative leaf and stem samples were subsequently collected and weighed. A sub
sample of leaf laminae was also collected and leaf area determined as previously
described. Leaf and stem samples were then oven dried at 65 QC for 48 hours before
being weighed dry and ground to pass a 0.5 mm screen. Ground tissues were analyzed
for Total N, P and K concentrations as previously described. Root biomass was not
sampled.
3 . 1 .4.3 Fruit Yield and Quality Measurements
One destructive fruit harvest was made 1 30 DAT to determine the effect of
fertilizer application rate on yield and quality parameters. Fruit harvests planned for
later in the season (including a final harvest at commercial maturity, > 95 % ripe) were
abandoned because unseasonably high summer rains resulted in a substantial incidence
76
of late blight and also increasingly large rot losses. At 1 3 0 DAT, all fruit within a
representative 3 m section of each split-plot was hand-harvested and weighed to
determine total yield. Fruit were sorted according to three categories, including
marketable, green, and culls (including pathological rots and limited-use fruit). As
expected, BER incidence in the field was negligible compared to aeroponics. Blight
affected fruit were not designated as culls, because its occurrence clearly overestimated
that typical of commercial production. The number and mass of fruit in each category
was recorded, and mean fruit mass calculated.
A sample of approximately 20-25 intact marketable fruit per split-plot was
collected for mineral analysis. Fruit were cut longitudinally, weighed fresh and then
oven dried at 65 °C before being weighed dry. Samples were subsequently ground to
pass through a 0.5 mm screen and analyzed for Total N, P and K as previously
described. An additional 20-25 intact marketable fruit per split-plot were juiced and
filtered (using cheese-cloth) before determining SSC by temperature-compensated
refractometer.
3. 1 . 5 Statistical Analysis
Experimental data were SUbjected to split-plot ANOV A tests using the SAS
GLM procedure. SAS regression procedures were used to test the significance of linear
and quadratic effects in describing plant responses to the quantitative fertilizer variable.
Relevant log and arc-sine transformations were made where appropriate. A series of
orthogonal contrasts were used to compare specific fertilizer treatments combinations.
Correlation analysis was used to report positive and negative linear relationships
between variables of interest.
3. 2 Results
3.2 . 1 Leaf Nutrient Sampling
3 .2 . 1 . 1 Nitrogen Concentration in YML Tissue
Seasonally, N concentration in YML tissue declined in all treatments (Table
3 .4). On no date was there a significant interaction between treatment and cultivar.
77
Table 3.4. Effect of fertilizer application rate on YML N concentration (%) at early bloom,
full bloom and early fruit ripening.
Fertilizer treatment
o (control) 0.5 1 .0 1 .5 2.0
Sufficiency levels Y
Treatment x cultivar interaction
Linear effect
Quadratic effect
Contrasts Control vs. applied fertilizer Optimal ( 1 .0) vs. Iow (0 + 0.5) Optimal ( 1 .0) vs. high ( 1 .5 + 2.0)
Cultivar Morse H225
p -value
N concentration (%) in YML tissue Z Early bloom Full bloom Early ripening
4 . 1 4.4 4.6 4.9 5.0
4.6 · 5.2
ns
***
***
4.6 4.6
ns
3.9 4.1 4.4 4.6 4.7
3.5 · 4.5
ns
*** ***
4.3 4.3
ns
3.4 3.7 4 .0 4.1 4.1
2. 7 · 3.8
ns
***
***
ns
3.7 3.9
Z early bloom, full bloom and early ripening ( 1 0 % of fruit showing red colour) corresponded to 55, 74 and 99 OAT, respectively
Y sufficiency levels from Hartz et al. (1 998) ns , ., **, *** non-significant at p < 0 . 10, or significant at p < 0.10 , 0.05 or 0.01 , respectively
Both linear and quadratic fertilizer effects were highly significant at early
bloom, full bloom and early fruit ripening; on each date, the quadratic effect had a
slightly higher coefficient of determination (r2; a measure of goodness of fit) value.
Across dates, N concentration in YML tissue began to plateau at the highest fertilizer
rate. The optimal fertilizer rate (1 .0x) generally resulted in tissue concentrations that
were intermediate to lower and higher application rates. On all three dates the no
fertilizer control had significantly lower tissue N values when compared directly against
grouped fertilizer application (i.e. all rates of fertilizer). The remaining contrasts that
were significant followed a similar trend; a greater level of fertilizer application resulted
in higher tissue N concentration. YML N concentration at early bloom was below
established sufficiency norms for both the no fertilizer control and O.5x treatments. On
the remaining dates, tissue N concentrations were interpreted as sufficient for plant
growth. Cultivar had no effect on YML N concentration during early bloom and full
78
bloom. However, late in the season 'H225 ' had significantly higher tissue N
concentration than 'Morse' , presumably related to differences in cultivar maturity (N is
typically redistributed from the leaf as fruit develop).
3 .2. 1 .2 Phosphorus Concentration in YML Tissue
There was no P fertilizer treatment during the field study. This reflected the
very high soil test P value at the experimental site (bicarbonate-extractable P was 80 mg
kg- I ) . As expected, there was no effect of N and K fertilizer application on YML P
levels . Across treatments, P concentration declined seasonally, averaging 0.45, 0.39
and 0.33 % during early bloom, full bloom and early fruit ripening, respectively. On all
dates, these values reflected tissue concentrations interpreted as sufficient. All contrasts
were not significant. YML P concentrations were essentially identical for 'Morse' and
'H225 ' .
3 .2 . 1 .3 Potassium Concentration in YML Tissue
Seasonally, K concentration in YML tissue declined in all treatments (Table
3 .5) . On no date was there a significant interaction between treatment and cultivar.
Both linear and quadratic fertilizer effects were highly significant at all growth stages;
across dates, the quadratic response had a higher r2 value. On all dates, K concentration
in YML tissue had a notable plateau at the higher application rates. The no fertilizer
control had consistently lower tissue K values when compared directly against grouped
fertilizer application. The remaining contrasts that were significant followed a similar
trend; a greater level of fertilizer application resulted in higher tissue K concentration.
YML K concentration at all growth stages appeared to be above established sufficiency
norms, even in the no fertilizer control. Cultivar had no effect on YML K concentration
during early bloom and full bloom. However, late in the season 'H225 ' had
significantly higher tissue K concentration than 'Morse' . As with N, this was
presumably related to differences in cultivar maturity.
79
Table 3.5. Effect of fertilizer application rate on YML K concentration (0/0) at early bloom,
ful l bloom and early fruit ripening.
K concentration (%) in YML tissue Z Fertilizer treatment Early bloom Full bloom Early ripening
o (control) 0.5 1 .0 1 .5 2 .0
Sufficiency levels Y
Treatment x cultivar interaction
Linear effect
Quadratic effect
Contrasts Control vs. applied fertilizer Optimal (1 .0) vs. Iow (0 + 0.5) Optimal ( 1 .0) vs. high ( 1 .5 + 2.0)
Cultivar Morse H225
p -value
4.3 4.7 4.8 4.9 4 .9
2.2 - 3.5
ns
***
ns
4.7 4 .7
ns
3.9 4 . 1 4.3 4.5 4.6
1.6 - 4. 1
ns
*** ***
ns
ns
4.2 4.4
ns
2.7 3.8 4.1 4.2 4.4
0.8 - 2.0
ns
***
ns
3 .6 4 .1
Z early bloom, full bloom and early ripening (10 % of fruit showing red colour) corresponded to 55, 74 and 99 OAT, respectively
Y sufficiency levels from Hartz et al. ( 1 998) ns , ** , *** non-significant at p < 0 . 1 0, or significant at p < 0.05 or 0.01 , respectively
3.2 .2 Destructive Whole-Plant B io mass Sampl ing
One destructive whole-plant sampling was made 55 DAT, corresponding
approximately to flowering and early fruit set. Most vegetative development was
thought to have occurred by this date; because of the determinate growth habit in
processing tomato, there is generally little new 'structural ' (leaf and stem) biomass
developed once fruit sinks establish. Based on the previous aeroponic study, it also
appeared that fruit yield was influenced most significantly by this period of early plant
growth (i .e. once the vegetative framework was formed and fruit number established,
yield outcomes were largely determined). As expected, there was no substantial fruit
biomass at 55 DAT.
No variable had a significant interaction between treatment and cultivar. Total
plant biomass and leaf biomass response to fertilizer application was significantly
80
described by both linear and quadratic effects (Table 3 .6); in both instances, the
quadratic response had higher r2 values. Leaf biomass began to plateau more distinctly
at the highest application rate than total biomass. For stem biomass, the linear response
had a slightly higher r2 value. Contrasts confirmed that fertilizer application increased
all three biomass measures when compared collectively against the control treatment.
Less than optimal fertility (no fertilizer control and 0.5x treatments) resulted in
significantly lower total plant and leaf biomass when compared directly against the
optimal treatment. There was no effect of fertilizer application rate on partitioning
between the leaf and stem, across treatments representing approximately 70 and 30 % of
total biomass at 55 DAT, respectively.
Table 3.6. Effect of fertilizer application rate on plant dry biomass (total, leaf and stem)
and leaf area at 55 OAT.
Total biomass leaf biomass Stem biomass leaf area
Fertilizer treatment (g planr1) (g planr1 ) (9 planr1 ) (cm2 planr1 )
0 (control) 61 43 1 9 8240
0.5 72 52 20 1 0057
1 .0 77 54 23 1 0477
1 .5 83 59 24 1 1 551
2 .0 87 61 26 1 1 960
Treatment x cultivar interaction
ns ns ns ns
Linear effect *** *** ** *** Quadratic effect *** *** ***
Contrasts
Control vs. applied fertilizer *** *** *** Optimal ( 1 .0) vs. Iow (0 + 0.5) * ns ns
Optimal ( 1 .0 ) vs. high ( 1 .5 + 2.0) ns ns ns ns
Cultivar Morse 69 49 20 9470
H225 83 58 25 1 1 444
p -value *** *** ***
ns , *, **, *** non-significant at p < 0 . 1 0, or significant at p < 0 . 1 0, 0.05 or 0.01 , respectively
The effect of cultivar on biomass measures (total, whole-leaf and stem) was
strongly significant; in all instances 'H225 ' had greater biomass than 'Morse' . As with
treatment, there was no substantial difference in relative biomass partitioning to the leaf
and stem between cultivars. Across cultivars, total plant biomass was positively
8 1
correlated with N and K concentration in YML tissue at 55 DAT (p < 0.0 1 , r = 0.57 and
0.56, respectively), but not P concentration.
Leaf area response to fertilizer application was also significantly described by
both linear and quadratic effects; the quadratic response had a higher r2 value, although
again this difference was small. Leaf area began to plateau at the highest application
rates. Fertilizer application increased leaf area when compared collectively against the
control treatment. The remaining contrasts were not significant. As with plant biomass,
the effect of cultivar on leaf area was highly significant; 'H225 ' had a larger leaf area
than 'Morse' . Leaf area was strongly correlated with leaf biomass (p < 0.01 , r = 0.89).
Leaf area was also positively correlated with N and K concentration in YML tissue at
55 DAT (p < 0.01 and 0.02, r = 0.65 and 0.43 , respectively), but not P concentration.
Leaf area index (LAI, the proportion of leaf area per unit area of individual plant space)
ranged from between 1 .8 for the no fertilizer control to 2.7 for 2.0x the optimal fertilizer
application rate; a higher index indicates improved vegetative cover, and may allow for
greater interception of solar radiation (and therefore increased plant productivity).
3.2 .3 Fruit Yield and Qual ity Measurements
Although the final harvest was not made at full commercial maturity (i . e. > 95
% ripe fruit), yield results did not appear to be greatly compromised. In a separate
maturity experiment (using a glasshouse solely as a rain-shelter) conducted parallel to
the field study, ripening rates (expressed as percent colour change per day) were
established for 'Morse' and 'H225' . Using these ripening rates (both were
approximately 2 .5 % colour change per day) an estimation of expected commercial
maturity was made for the control and 2.0x border treatments; 'Morse' would have
required an extra 3- 10 days to reach a 95 % ripe-stage of maturity ( 1 34-1 4 1 DAT) ,
while 'H225' would have required an extra 1 1 -20 days ( 1 42- 1 5 1 DAT). Importantly,
fruit set within the final six weeks do not typically mature by harvest; therefore, all fruit
that would have contributed to commercial yields were established well in advance of
the destructive fruit harvest at 1 30 DAT.
The effect of fertilizer application rate and cultivar on tomato yield and fruit
quality is summarized in Table 3.7. No variable had a significant interaction between
treatment and cultivar.
82
Table 3.7. Effect of fertilizer application rate on tomato yield and fruit at 1 30 OAT.
Total Marketable Soluble Brix Adjusted % Ripe y Fertilizer treatment yield yield solids �eld Brix �eld z
(Mg ha-1) (Mg ha-1) (OBrix) (Mg Brix solids ha-1 )
0 (control) 68 42 4.09 1 .71 2 .50 77
0.5 82 47 4 . 1 1 1 .93 3 .00 73
1 .0 92 48 4.24 2.04 3 .53 63
1 .5 1 00 51 4.42 2.27 3.98 62
2.0 1 03 5 1 4.42 2.27 4 .09 59
Treatment x cultivar interaction
ns ns ns ns ns ns
Linear effect *** *** *** *** Quadratic effect *** ** ** *** *** ***
Contrasts
Control vs. applied fertilizer *** .. ** * *** ns Optimal (1 .0) vs. Iow (0 + 0.5) * ns ns ns ns Optimal ( 1 .0) vs. high ( 1 .5 + 2.0) ns ns .. ns ns ns
Cuftivar Morse 78 51 4.34 2 .22 3.04 79
H225 1 0 1 45 4. 1 7 1 .86 3.79 55
p -value *** ** ** *** *** ***
Z adjusted Brix yield based on 90 % of total yield considered marketable Y ripe fruit included both marketable and cull constituents ns , *, **, *** non-significant at p < 0 . 10 , or significant at p < 0. 1 0, 0.05 or 0.01 , respectively
3 .2.3 . 1 Total and Marketable Yield
Total and marketable yield response to fertilizer application was significantly
described by both linear and quadratic effects (Table 3 .7). However, in both instances,
the quadratic response had a higher � value. Total yield began to plateau more
distinctly at the highest fertilizer application rates than marketable yield, presumably
related to the fact that the final harvest occurred before full maturity (across treatments,
marketable yield only averaged 54 % of total yield at 1 30 DAT). Total fruit yield was
low « 70 Mg ha-I) in control plots that received no fertilizer. Fertilizer application
significantly increased total fruit yield when compared collectively against the control
treatment. Less than optimal fertility (control and 0.5x treatments) resulted in lower
total yield when compared directly against the optimal treatment ( I .Ox). Higher total
fruit yield was more strongly correlated to total fruit number than mean fruit mass (p <
0.0 1 , r = 0.92 and 0 .79, respectively) . The effect of fertilizer application on cultivar
83
was also significant; total yield was greatest for 'H225 ' , reflecting both a larger number
of fruit per plant and larger mean fruit mass. Significant quadratic regressions (p <
0.0 1 ) were fitted to total yield against fertilizer application for both 'Morse' and ' H225'
(Fig. 3 .2).
1 40�----------------------------------------------�
1 20
'7 m .r::. 1 00
r2 = 0.6� _ - - - -- -- ---- - - - -I -
- - - - - - - -- -- -
- - - - - - - -I J;- - ---
.. _-- - I - - - -y 0> � "0 ID 's>. 80 -co ..... 0
... . . . . .. I·· ·· ·· ·· · · · ·· · ·· · ·· · · ·I .. . . · · · · · ·· · · · · · · I r2 = 0.66
I-
60
40�_,,----�----r-1--�----,,-----r----.-,--�----.-� 0.0 0.5 1 .0 1 .5 2.0
Fertilizer treatment (x optimal rate)
Fig. 3.2. Effect of fertilizer application rate on the total yield of 'Morse' and 'H225' at 130 OAT. Fertilizer treatments represented 0 (control), 0.5, 1 .0, 1 . 5 and 2.0x the optimal fertilizer rate (268 kg N ha·1 and 462 kg K ha·1). Treatment means are given ± standard error. Cultivars were 'Morse' ('\7) and 'H225' (0). From replicated data, the equation of the fitted quadratic regression line ( ... ) for 'Morse' was y = 60 + 23.2x - 4.0"( and for 'H225' y = 75 + 40.8x - 1 0.3.,(.
Total plant dry biomass at 55 DAT (comparable with full vegetative growth)
was positively correlated with total fruit yield (p < 0.0 1 , r = 0.64) . Leaf area index at 55
DAT was also positively correlated with total yield (p < 0.0 1 , r = 0.83); an index above
approximately 2 .3 was generally associated with fruit yields greater than 90 Mg ha- ' (an
acceptable commercial benchmark). Total yield was positively correlated with N
concentration in YML tissue during early bloom (p < 0.0 1 , r = 0.63); tissue levels 2: 4.6
% N (the minimum sufficiency value for this phase of growth) were associated with
fruit yields above 90 Mg ha-i . YML N concentration at full bloom and early fruit
ripening was also associated with final fruit yield (r = 0.59 and 0.74, respectively);
however, these correlations did not appear noteworthy, because both samplings
occurred after most vegetative growth had stopped (it was vegetative growth during the
84
initial period that detennined fruit number) . Although YML K concentration at early
bloom, full bloom and early fruit ripening was significantly correlated with total yield
(p < 0.05, r = 0.4 1 , 0.40 and 0.54, respectively), in all instances tissue K levels were
above established sufficiency nonns. YML P concentration was not correlated with
yield outcomes.
The effect of fertilizer application on fruit culls was not significant. Across
treatments, culls averaged approximately 1 0 %, although in some individual plots they
were as high as 30 % (data not shown). Culled fruit were almost exclusively due to
pathological rots, in particular water mo1d; this apparently reflected the very wet
conditions in the field. Although not significant, culls were marginally higher for
'Morse' compared to 'H225' ( 1 2 and 9 %, respectively); this appeared related to the
earlier maturity of 'Morse' , as ripe fruit are more susceptible to rots (Strand, 1 998) .
3 .2.3 .2 Fruit Soluble Solids
Across treatments, SSC of marketable fruit was low, averaging < 4.5 °Brix
(Table 3 .7) ; as previously suggested, an acceptable target for fruit grown commercially
is > 5 .0 °Brix. Fruit SSC response to fertilizer application was significantly described
by both linear and quadratic effects, although neither was superior to the other.
Fertilizer application significantly improved fruit SSC when compared collectively
against the control treatment. Similarly, above optimal fertility ( 1 .5 and 2.0x
treatments) resulted in higher SSC when contrasted directly to the optimal treatment. In both instances, the improvement in SSC was only modest.
Fruit SSC was significantly higher for 'Morse' , although this difference was also
comparatively small « 0.2 °Brix) . Fruit SSC was not significantly correlated to total
yield; high yields can cause a 'dilution' of fruit SSC. Leaf area index (as an indicator of
assimilate supply in plants) at 55 DAT was also not significantly correlated to
subsequent SSC potential; this was despite a numerical trend towards higher SSC with
larger early-season LAI values. Fruit SSC was not correlated with K concentration in
fruit at 1 30 DAT.
85
3 .2.3 .3 Brix Yield and Adjusted Brix Yield
Across treatments Brix yields were low, averaging approximately 2 Mg Brix
solids ha- I (Table 3 .7); as previously suggested the commercial norm is at least 5 Mg
Brix solids ha- I . This not only reflected the earlier than ideal harvest date (and therefore
a low percent of marketable fruit), but also poor fruit SSC levels « 4.5 °Brix). Despite
this, Brix yield response to fertilizer application rate was still significantly described by
linear and quadratic effects; the quadratic response had a slightly higher r2 value. Brix
yields appeared to plateau at the highest application rates. Fertilizer application
significantly improved Brix yield outcomes when compared collectively against the
control treatment. The remaining contrasts were not significant. Brix yield was more
closely correlated with marketable yield than with fruit SSC (p < 0.0 1 , r = 0.97 and
0.6 1 , respectively). Brix yields were significantly higher for 'Morse' rather than
'H225 ' ; however, this observation appeared to be overstated, as 'H225 ' had
significantly higher total fruit yields but delayed maturity.
Because the final harvest occurred before fruit were fully ripe, an adjustment of
Brix yield was made to reflect a typical field maturity; marketable yield was considered
as 90 % of the total yield measured at 1 30 DAT (assuming 5 % greens and 5 % culls).
Although higher, adjusted Brix yields still remained below 5 Mg Brix solids ha-I .
Similar statistical patterns were found to the original data. The effect of cultivar
remained significant, although as expected, with the adjustment ' H225' had greater Brix
yield outcomes than 'Morse' (because of higher total yield).
3 .2.3 .4 Percent Ripe Yield
Percent ripeness is an indicator of crop maturity; a lower value reflects fewer
ripe fruit. As noted, the typical crop maturity at which commercial fields are harvested
is approximately 95 % ripe fruit; this compared to only 67 % across treatments in the
current study. Plans for destructive sampling later in the season were abandoned
because of heavy summer rainfall and increasing late-blight pressure and fruit rot
concerns. Ripe fruit included those that were either marketable or culls (culls were
overwhelmingly ripe fruit). At 1 30 DAT, the percent of fruit that were ripe in response
to fertilizer application rate was significantly described by linear and quadratic effects
(Table 3 .7); the quadratic response had a marginally higher r2 value. The highest
86
fertilizer application rates resulted in the lowest percent of ripe fruit. However, all
contrasts were not significant.
The percent of ripe fruit was significantly higher for 'Morse' rather than 'H225 ' .
This difference was consistent with the growth patterns of the two cultivars ( 'Morse' i s
early-maturing, while 'H225 ' i s mid-maturing). Percent ripe fruit was negatively
correlated with YML N concentration at early boom, full bloom and early fruit ripening
(p < 0.05, r = -0.39, -0.48 and -0.63, respectively). Although YML K concentration was
also negatively correlated with percent ripeness on these dates, coefficients were
consistently weaker.
3 .2 .3 .5 Fruit Nutrient Concentrations
There was no interaction between treatment and cultivar for any fruit nutrient
measured (Table 3 .8).
Table 3.S. Effect of fertilizer application rate on the concentration of N, P and K (%) in
fruit at 130 OAT.
Fertilizer treatment
o (control) 0.5 1 .0 1 .5 2.0
Treatment x cultivar interaction
Linear effect
Quadratic effect
Contrasts
Control vs. applied fertilizer Optimal ( 1 .0) vs. Iow (O + 0.5) Optimal ( 1 .0) vs. high (1 .5 + 2.0)
Cultivar Morse H225
p -value
Fruit tissue concentrations (%)
N P
3.1 0.48 3.3 0.50 3.4 0.50 3.5 0.49 3.5 0.51
ns ns
*** ns ** ns
** ns
ns ns
ns ns
3.5 0 .48 3.2 0.51
*** ns
ns , * , **, *** non-significant at p < 0 . 10 , or Significant at p < 0 . 10, 0.05 or 0.01 , respectively
87
K
4.4 5 . 1 5.3 5.4 5.6
ns
*** ***
*** ***
5.2 5 . 1
*
The effect of fertilizer application rate on both fruit N and K concentration was
significantly described by linear and quadratic effects; in both instances, the quadratic
response had a higher r2 value. There was no effect ofN and K fertilizer application on
tissue P levels in fruit; across treatments, P concentration averaged approximately 0.50
%. Collectively, fertilizer application significantly improved fruit N and K concentrations when compared against the control treatment. Less than- and above�
optimal fertility resulted in lower and higher fruit K concentrations when compared
separately against the optimal treatment, respectively. All other contrasts were not
significant. The concentration of N and K in fruit tissue was significantly greater in
'Morse' compared to 'H225' ; this difference was however comparatively small. There
was no effect of cultivar on fruit P concentration.
3 .2.3 .6 Nutrient Removal and Fertilizer Efficiency
Nutrient removal rates in the harvested fruit were calculated using an adjusted
marketable yield (90 % of total yield at 1 30 DAT). Because the final harvest occurred
before full maturity, this standardized all treatments to a comparable level of fruit
ripeness. Based on adjusted marketable yields, the effect of fertilizer application rate on
the removal of N, P and K in fruit was significantly described by linear and quadratic
effects (Table 3 .9); in all instances, the quadratic response had a marginally higher r2
value. For all significant contrasts, a higher level of fertilizer application resulted in
greater removal of N, P or K in marketable fruit. The effect of cultivar was also
significant for each nutrient; in all instances, 'H225 ' had the greatest rate of removal
(consistent with higher overall yields for this cultivar). Removal rates for N, P and K
were strongly correlated with adjusted marketable yields (p < 0.0 1 , r = 0.93, 0.89 and
0.94, respectively); although the concentration of each nutrient in fruit was correlated
with removal, coefficients were weaker (r = 0.33, 0.67 and 0.56, respectively).
Fertilizer efficiency (Fe) was estimated only for N and K treatments; P was not
applied differentially during the experiment. The Fe value was calculated as the mass of
nutrient in marketable fruit (less that removed by the control which received no
fertilizer application) expressed as a percent of the mass of fertilizer applied. To
standardise for differences in maturity, the adjusted marketable yields for each
treatment were used. For N, Fe values ranged from 22 - 1 2 % across the 0.5 and 2 .0x
treatments, respectively; recovery was also greater in 'H225' than 'Morse' (consistent
88
with higher fruit yields). Similar trends were observed for K; Fe values ranged from 20
- 13 % across the 0.5 and 2.0x treatments, respectively, with recovery also being greater
in 'H225 ' . Efficiency values did not account for the mass of applied fertilizer that was
essential in establishing the vegetative framework, but that was not redistributed to the
fruit (i .e. that which is eventually returned to the soil as crop residue).
Table 3.9. Effect of fertil izer application rate on removal of N, P and K in marketable fruit
at 1 30 OAT.
Nutrient removal (kg ha·1) in marketable fruit at 1 30 DAT z
Fertilizer treatment N P K
o (control) 0.5 1 .0 1 .5 2.0
Treatment x cultivar interaction
Linear effect
Quadratic effect
Contrasts
Control vs. applied fertilizer Optimal ( 1 .0) vs. Iow (0 + 0.5) Optimal (1 .0) vs. high ( 1 .5 + 2.0)
Cultivar Morse H225
p -value
84 1 09 1 30 1 40 1 46
ns
*** ***
***
ns
1 1 2 1 31 ***
1 3 1 7 1 9 20 22
ns
*** ***
** ns
ns
1 5 21
***
Z calculated using adjusted marketable yields which were 90 % of total fruit yield ns , **, *** non-significant at p < 0. 10 , or significant at p < 0.05 or 0.0 1 , respectively
3.3 Discussion
1 20 1 67 200 2 1 8 236
ns
*** ***
*** *** ns
1 66 2 1 0 ***
3.3 . 1 Effect of Fertil izer Appl ication on Leaf Nutrient Concentrations
3 .3 . 1 . 1 Nitrogen Concentration
Leaf N concentrations declined seasonally as assimilates were redistributed to
developing fruit. This trend was similar to that observed under aeroponics, and is also
89
consistent with that commonly reported in the literature (Colla et aI. , 200 1 ; Hartz et al. ,
1 998; Lorenz and Tyler, 1 983 ; Tei et al. , 2002). Optimal fertility and above resulted in
tissue N values interpreted as sufficient for high yield (Hartz et al. , 1 998). In plants
receiving above optimal fertility, tissue N concentrations were not supra-optimal despite
very large N applications of up to 540 kg N ha-1 (as much as double typical commercial
applications; Hartz and Miyao, 1 997); this suggested plant demand was not as high as
these rates supplied. Tissue N concentration was low during early bloom in treatments
receiving less than optimal fertility; tissue values appeared to be mildly deficient, and
were comparable to concentrations found under aeroponics to reduce plant growth and
fruit yields. However, late-season tissue N concentration in treatments receiving less
than optimal fertility was not below sufficiency values. This suggested that N application after most active fruit growth has stopped was unnecessary.
3 .3 . 1 .2 Phosphorus Concentration
All treatments resulted in adequate tissue P status compared to established
sufficiency norms (Hartz et al. , 1 998). This apparently reflected the high soil test P
value at the experimental site (80 mg kg- I P); a bicarbonate-extractable value of > 1 5
mg kg- I P is generally interpreted as sufficient for maximum yield (Hartz and Miyao,
1 997). Leaf P concentration declined seasonally, as P was redistributed to fruit. This
was in contrast to aeroponics (where tissue P levels remained high throughout the season), but consistent with prior field observations (Hartz et al. , 1 998; Hochmuth et
al. , 1 999; Lorenz and Tyler, 1 983). Tissue P concentrations were also considerably
lower than found under aeroponics (0.7-0.8 % P), even at what was considered to be a
high soil test P level. This confirmed that there was luxury uptake of P in the
aeroponics experiment. As previously noted, while excessive P fertilization generally
does not cause serious agronomic problems, it can be a significant contributor to P
pollution of surface waterways (Pote et al. , 1 996).
3 .3 . 1 .3 Potassium Concentration
All treatments resulted in adequate to high tissue K status compared to
established sufficiency norms (Hartz et al. , 1 998). Seasonal decline was observed in all
treatments, even those supplying large amounts of fertilizer K; this reflected
90
redistribution of K to the fruit, and was consistent with prior studies (Besford and Maw,
1 975 ; Hartz et al. , 1999, 2005; Pan et aI. , 1 999; Widders and Lorenz, 1 982). It also
suggested that K uptake by the roots could not match fruit demand during this active
period of growth; fruit are the major sink for absorbed K (Adams, 1 986). Although K
application during fruit ripening resulted in greater leaf K concentrations, such
applications were unnecessary for maximum yield and fruit quality. Even the control
treatment which applied no K fertilizer resulted in tissue concentrations above
sufficiency values. This suggested that K application after most active fruit growth has
stopped was unnecessary. Field results confinned that the high late-season tissue K
concentrations (> 5.0 %) under aeroponics represented lUXUry levels; K uptake in these
conditions therefore over-estimated actual plant demand.
3.3 .2 Effect of Fertil izer Application on Whole-Plant Biomass
Plants receiving optimal fertility and above established the greatest total biomass
and leaf area by flowering. The N status of plants was most important in accounting for
these differences; plants with low biomass and small leaf area (the no fertilizer control
and the 0.5x treatments) had leaf tissue N concentration below sufficiency nonns at
early bloom. The role of N in vegetative growth has been previously documented;
when deficient, canopy development can be limited (Cavero et al. , 1 999; Colla et al. ,
1 999, 200 1 ; Pan et al. , 1 999) and radiation interception and photosynthetic activity
reduced (Greenwood, 2001 ; Sinclair and Horie, 1989; Tei et ai. , 2002). The P and K status of plants were adequate for all treatments. The positive correlation between
YML K concentration and plant biomass and leaf area appeared then to be largely
coincidental, apparently reflecting increased K availability with increased N supply.
These findings were consistent to those made under aeroponics, and further
reinforced the importance of establishing a strong vegetative framework early in the
season for subsequent flowering and fruit development. Once vegetative growth stops
and fruit number is detennined, yield outcomes are largely set (i.e. only fruit size can be
positively influenced). These results were consistent to commercial fertilizer trials
conducted in the Hawkes Bay during 1 994-95 (the major production region of N.Z.);
conclusions from this work advised growers to maximize early growth in order to
achieve a large vegetative framework for flowering (BASF, unpublished data).
9 1
3.3.3 Effect of Fertil izer Application on Fruit Yield and Qual ity
3.3 .3 . 1 Total and Marketable Fruit Yield
High total fruit yield was achieved with treatments that supplied optimal fertility
and above. Although application rates above optimal resulted in the numerically greater
fruit yields, these were not statistically different to those achieved with the optimal
treatment. Higher yield was due primarily to greater fruit number rather than mean fruit
mass. Improved fruit number was associated with the larger vegetative frameworks
established early in the season; this apparently reflected an improved assimilate
capacity in plants. These findings were consistent to those made under aeroponics, and
were also related to early-season N status in plants rather than to P or K.
Marketable yields were not maximized for any treatment; this reflected an early
harvest date due to an increasing incidence of plant disease and fruit rots. Above
optimal fertility resulted in the lowest percent of total yield that was marketable. This
was linked to high tissue N status which delayed maturity. Similar observations have
been made by May and Gonzales ( 1 994) and Nicklow and Downes ( 1 97 1 ); both found
larger green fruit yields within increasing N application rates. However, there was no
effect of high fertility on fruit culls. If all treatments had been grown to commercial
maturity, marketable yields would have been greatest in plants receiving optimal
fertility and above. As expected, the incidence of BER was almost non-existent in the
field compared to levels found under glasshouse aeroponic culture. This apparently
reflected high moisture supply in the soil during fruit growth and development;
intermittent water stress during this critical period is the most common cause of BER in
the field.
Due to heavy summer rainfall and light-textured soil type, the likelihood of N
(and possibly K) leaching out of the active root zone was high; plants were already
receiving full replacement of ET by irrigation water (which was necessary to fertigate
on nutrient treatments). Because fruit yields were comparable at application rates of
optimal fertility and above, optimal fertility itself would most likely have been
sufficient to maximize growth and subsequent yields under normal field conditions (i.e.
a drier season). This was supported by the fact that typical grower N applications for
high yields are considerably less than that for the above optimal treatments; the 2.0x
92
treatment supplied approximately 540 kg N ha-I , while maximum grower applications
do not commonly exceed 280 kg N ha- I .
Fertilizer applications made during fruit ripening were not effective in
increasing yield, as mean fruit mass was not greatly improved. Maximum yields
therefore appeared attainable at rates equivalent to optimal plant uptake as determined
for only the initial two periods of growth under aeroponics (vegetative and fruit
development); at a density of 22,222 plants ha- I this was equal to approximately 2 1 0
and 3 1 0 kg ha- I of N and K, respectively. This type of N application was consistent
with current grower applications ( 140-280 kg N ha-I ; Hartz and Miyao, 1 997), although
is still greater than estimates suggested by Krusekopf et al. (2002). Although K
application amount appeared considerably higher than current grower levels « 1 1 0 kg
K ha- I ; Hartz and Miyao, 1 997), it was comparable to a 'typical' plant requirement of
approximately 300 kg K ha-I as suggested by Widders and Lorenz ( 1 979); the apparent
disparity between grower supply and plant demand is balanced by high K supply from
the soil.
3 .3 .3 .2 Fruit Soluble Solids
Across treatments, low fruit SSC reflected high soil moisture availability during
ripening and also a final harvest before full maturity; both factors can substantially
reduce SSC outcomes. These observations were confirmed in a separate study that ran
concurrently to the field trial (production details are provided by Nichols et al. , 200 1 );
using a glasshouse solely as a rain-shelter, fruit SSC at commercial maturity was 5 .6
and 5.0 °Brix for 'Morse' and 'H225' , respectively (in the field, fruit SSC was
consistently < 4.5 °Brix). This confirmed that low SSC was not inherently related to
these two cultivars, but to the production environment in which they were grown.
Higher fertility resulted in greater fruit SSC. This improvement may have been
associated with the larger plant canopies established with higher fertilizer rates; larger
canopies often have improved assimilate capacity (SSC is in part reflective of dry
matter content in fruit resulting from assimilate import). Surprisingly, there was not a
stronger correlation between early-season LAI values and fruit SSC; this may have
reflected variability in the leaf area measure which was determined on only a sub
sample of tissue (LAI was subsequently calculated from leaf area).
93
Fruit tissue K concentration did not positively affect fruit SSc. Combined with
the weak effect under aeroponics, these observations were largely consistent with those
made by Hartz et al. (200 1 , 2005) ; across trials and seasons, they found no relationship
between fruit SSC and K fertility (even when fruit K concentrations increased). The
results of Lachover ( 1 972) suggesting a positive relationship between fruit SSC and K
status therefore appeared to be an anomaly, perhaps related to production in small pots
where rooting was restricted. Hartz et al. (2005) suggested such an environment may
be substantially different to the field setting. Application of K fertilizer during fruit
ripening was therefore neither a reliable nor cost effective means to favourably
influence SSC, and did not appear to offer any further benefit beyond appropriate late
season water management; considerably higher fruit SSC levels were achieved without
excessive K application but where late-season soil moisture stress was imposed by rain
shelter.
3 .3 .3 .3 Brix Yield and Adjusted Brix Yield
Low Brix yields reflected both modest-low fruit yield and poor SSC; even when
total yield was adjusted to a standardized commercial maturity (based on 90 % of total
yield considered marketable), adjusted Brix yields remained < 5 Mg Brix solids ha- I
(the commercial norm). These observations appeared to primarily reflect the growing
environment in Palmerston North during the 1 999-2000 season; heavy rainfall
combined with regular irrigation provided the ideal environment for disease, which
appeared to limit yields. Similarly, high moisture availability during fruit ripening
certainly appeared to reduce Ssc. Given the positive effect of at least optimal fertility
on total yield and fruit SSC, Brix yield was also maximized for these treatments.
3 .3 .3 .4 Fruit Mineral Concentrations
The concentration ofN, P and K in fruit reflected the overall availability of each
nutrient in the plant, but otherwise appeared unimportant. Although fruit K
concentration was higher with increased K fertilizer rates, there was no improvement in
fruit SSC detected.
94
3 .3 .3 .5 Fruit Nutrient Removal and Fertilizer Efficiency
Nutrient removal in the harvested marketable product was closely associated
with fruit yields; higher yields resulted in greater removal of N, P and K. However, the
majority of applied N and K were not removed in the harvested fruit, and either
remained in the soil, was leached or returned to the soil as crop residue. Fertilizer
efficiency was poor, less than 25 % of applied N and K for all fertilizer treatments. The
very low efficiency at above optimal fertility confirmed that such applications were not
only unjustified, but extremely wasteful; these applications exceeded plant demand
under normal field conditions. Despite poor efficiency, fertigation technology still
offered the potential to closely match plant demand with supply compared to
conventional pre-plant and side-dressing fertilization techniques.
95
CHAPTER 4: 2001 and 2002 Irrigation Management Trials
Irrigation experiments were conducted at the Vegetable Crops Field
Headquarters (long. 1 2 1 0 47 ' W; lat. 380 32 ' N), University of California, Davis,
California during the summer of 200 1 and 2002. The aim was to evaluate the effects of
irrigation management during fruit ripening on yield and SSC of drip-irrigated
processing tomatoes; water deficits prior to ripening (i.e. vegetative and fruit
development) can dramatically reduce fruit set, decrease fruit size, and substantially
increase the incidence of fruit rots, and should therefore be avoided.
4.1 Materials and Methods
4. 1 . 1 Experimental Site
Davis has a Mediterranean-type climate with warm, dry summers and cool,
moist winters (Andrews, 1 972). No rainfall was received in either season during the
final 60 days preharvest. Daily reference evapotranspiration (ET 0) and air and soil
temperature were collated from the nearest computerized weather station of the
California Irrigation Management Information System (CIMIS) network. This weather
station was less than 1 km from the experimental sites. A full summary of climate data
for both seasons is provided in Appendix Ill.
Soil type for the 2001 study was a Reiff loam (coarse-loamy, mixed, non-acid,
thermic Typic Xerorthent) and in 2002 a Yolo loam (fine-silty, mixed, non-acid, thermic
Typic Xerorthent). Both soils were formed from alluvial fans, characterised with high
natural fertility and moderate permeability (Andrews, 1 972). The available water
holding capacity in the top 1 .5 m depth was between 2 1 5-250 mm for the Reiff soil and
229-279 mm for the Yolo soil (Andrews, 1 972).
4 . 1 .2 Production Detai ls
Standard commercial bed preparation techniques were made in the spring of
each season. A multi-row lister was used to form 1 .5 m-wide raised soil beds. A single
line of high-flow drip tape was buried 20 cm deep in the centre of the bed; spacing
96
between emitters was 30 cm. Preplant fertilizer was banded in both years; additional
side-dressed N was fertigated in the irrigation water as UN-32 (urea ammonium nitrate
solution). Seasonal application rates are summarized in Table 4. 1 .
Table 4.1 . Fertilizer application rates of N , P and K, 2001 and 2002.
Applied fertilizer (kg ha·1)
Year Nitrogen Phosphorus Potassium
Preplant Fertigated Z Prep/ant Prep/ant
2001 18 1 70 50 14 2002 31 1 70 1 9 24
Z applied as UN-32. split evenly over 5 (2001 ) or 6 (2002) weekly fertigations during plant development
The hybrid cultivar 'HaUey' was used in both years (Orsetti Seed Inc., Hollister,
CA); 'Halley' was selected because it was a dominant industry standard in California at
the time. Seed was sown and seedlings raised in a greenhouse until planting.
Transplants were hand-planted into single rows with in-row spacing of 30 cm; this was
equivalent to a density of approximately 22,000 plants ha' l (Fig. 4. 1 ). Sprinkler
irrigation was used to establish the transplants in both years; it also provided a full
moisture profile (field capacity) at the beginning of the season. Drip irrigation began on
either May 20th (2001) and June 3rd (2002). Important production and phenological
dates are summarized in Table 4.2.
Table 4.2. Production and phenological dates, 2001 and 2002.
Year
2001 2002
Sown
February 1 5 February 25
Transplanted First flowering
April 1 3 April 24
May 06 May 16
Phenological stage
First fruit set
May 20 June 03
First ripe fruit
July 01 July 1 3
In both years, herbicide control consisted o f pre-emergent application with
Devrinol 2E® (United Phosphorus Ltd.) at 2 .7 kg a.i ha' l , post-plant application with
Shadeout 25DF® (DuPont Chemical Co.) at 0.035 kg a.i ha' l , and early season lay-by
incorporation with Treflan 5E® (Dow Elanco Co.) at 1 .3 kg a.i ha'l . Insecticide control
for tomato russet mite (Aculops /ycopersici), potato aphid (Macrosiphum euphorbiae),
tomato fruit worm (Helicoverpa zea) and stink-bug (Eushistus conspersus) consisted of
97
multiple mid- and late-season applications with Thiodan® (Universal Crop Protection
Alliance Co.) at 1 . 1 kg a.i ha- I , Warrior T® (Syngenta Crop Protection, Inc.) at 0.34 kg
a.i ha- I and Ben-SuI 85® dusting sulphur (Wilbur-Ellis Co.) at 8 .5 kg a.i ha-I .
Fig. 4.1 . Early-season vegetative growth, 2001 .
4. 1 .3 Experimental Design
Seven irrigation treatments including various cutoff and cutback approaches
were evaluated in 200 1 to provide a full range of soil moisture deficits during fruit
ripening (Table 4.3). Experimental design was randomized complete block, with five
replications. Individual plots were three beds wide, each 1 5 m in length. Only the
centre bed was used for measurements, with the two outer beds designated as border
rows (in case of lateral water movement between treatments). A smaller subset of four
treatments was evaluated in 2002 (Table 4.3). Experimental design was again
randomized complete block, but with six replications. Each plot comprised one bed, 9
m in length; border rows were eliminated in 2002 because prior observations at this site
had confirmed no lateral water movement. The control treatment (full irrigation cutoff
20 days preharvest) used in both seasons represented the standard commercial practice
under drip. All treatments were designed to primarily target fruit ripening (water stress
prior to ripening can reduce fruit yields).
98
Table 4.3. Description of irrigation treatments, 2001 and 2002.
Treatment schedule
2001 Full cutoff implemented 20 days preharvest (control)
Full cutoff implemented 30 days preharvest Ful l cutoff implemented 40 days preharvest Full cutoff implemented 60 days preharvest (severe stress)
Cutback to 50 % of ETa implemented 40 days preharvest (full cutoff at 20 days) Cutback to 25 % of ETa implemented 40 days pre harvest (full cutoff at 20 days) Cutback to 25 % of ETa implemented 60 days preharvest (full cutoff at 20 days)
2002 Full cutoff implemented 20 days preharvest (control)
Full cutoff implemented 30 days preharvest Ful l cutoff implemented 50 days preharvest (severe stress)
Cutback to 25 % of ETa implemented 50 days preharvest (full cutoff at 20 days)
Full irrigation (I; Eq. 4. 1 ) was determined by multiplying daily ET 0, the crop
coefficient (Kc value x plant canopy cover), and a system inefficiency factor. Reference
ETo was calculated automatically by the CIMIS network using a modified-Penman
equation and hourly weather data (Goldhamer and Snyder, 1 989); closely clipped,
actively growing grass was the reference crop. The Kc value used for processing tomato
was 1 . 1 (Phene et al. , 1 986). Plant canopy cover was estimated using a ruler placed
laterally across the bed, and calculated as the proportion of 1 .5 m soil bed covered by
leaf (Hartz, 1 996); multiple measures were made to account for within field variability.
System inefficiency was set at 1 . 1 5 , a standard grower correction in drip-irrigated fields
(Hartz, 1 996); this value was higher than would have been expected in the
comparatively small plots lengths used for these experiments, where uniformity and
pressure loss were unlikely to be problematic.
Where:
I = full replacement irrigation amount (mm)
ET a = reference evapotranspiration (mm)
Ko = crop constant
p cc = plant canopy cover
Sf = system inefficiency factor
99
(Eq. 4. 1)
Water was applied three times per week and the amount verified by flow meters.
All treatments received 1 00 % of the calculated irrigation volume until each respective
cutoff/cutback was imposed. In both years, water supply to cutback treatments was
fully terminated at the same time as the control treatment (20 days preharvest). The
harvest date used to determine when to implement treatments was estimated using the
modified phenological growth model of Zalom and Wilson ( 1 999).
4. 1 . 4 Data Collection
4 . 1 .4. 1 Plant Canopy Cover Measurements
An infrared digital camera (Dycam Inc., Lost Hills, CA) was used to calculate
percent canopy cover; imagery was not related to P cc measurements made to determine
irrigation amount. The camera utilized spectral reflectance properties in the red and
near-infrared range to identify photosynthetically-active tissue in frame. Measurements
were taken weekly beginning seven weeks preharvest in 200 1 . This was approximately
1 0 days following the implementation of the 60 d treatments. A 2 m representative
section of row from the centre bed was imaged. On each date, four replications were
sampled, from four treatments selected to encompass the range of water stresses
imposed (60 d cutoff, 60 d cutback to 25 % of ET 0, 40 d cutoff, and the 20 d control) .
Plant canopy cover was not measured in 2002.
4. 1 .4.2 Plant Water Status
A pressure chamber instrument (Model 6 1 0, PMS Instruments, Corvallis, OR)
with a 4.0 MPa operational range was used to evaluate plant water potential ('V).
Measurements were taken weekly starting either seven (2001 ) or six (2002) weeks
preharvest, on a day when irrigation was not scheduled. Five replications per treatment
were sampled in both years. All samples were collected from healthy, recently-matured
leaves.
There has been little commercial application of 'V as an irrigation management
tool; the practical restrictions of predawn determination (only a few fields can be
monitored before sunrise) and the large variability associated with sampling exposed
leaves in the early afternoon have limited its use. To investigate the effect of sampling
1 00
technique on '" a comparison was made between leaf samples that had either been
bagged for a period of time prior to sampling or those that were exposed at the time of
sampling. Covering leaves with a reflective water-impervious bag allows equilibrium
to be reached between the leaf and stem (by stopping transpiration), and can therefore
improve the reliability of '" measures (Fulton et al. , 200 1 ; McCutchan and Shackel,
1 992).
Plant water status measured on bagged leaves represented stem water potential
("'Stem), while measurements made on exposed leaves (the traditional approach)
represented leaf water potential ("'Leaf). The effect of bagging technique was
investigated at both 0700 HR and 1300 HR; it was thought that sampling "'Stem at 0700 HR
would yield similar results to conventional predawn "'Leaf, whereas the 1 300 HR measure
was timed to coincide with peak solar load and highest plant water demand (and also
the greatest variability in exposed leaves). For the purpose of examining sampling
technique, matching samples of bagged and exposed leaves were only collected from
the control treatment at 0700 HR and the two border treatments at 1 300 HR (control and
most severe cutoff treatment in each year).
At 0700 HR, "'Stem was measured on leaves that had been bagged the previous
night (to allow adequate equilibrium before sampling). In both years, a conventional
predawn (0500 HR) measurement of "'Leaf was made to determine if 0700 HR sampling of
"'Stem was a reliable indicator of predawn plant water potential as predicted. This
measure was made either 14 (2001) or 27 (2002) days preharvest. On all other dates,
0700 HR bagged and exposed leaf samples were collected at the same time. At 1 300 HR, "'Stem was measured on leaves that had been bagged mid-morning ( 1 1 00 HR), allowing
an equilibrium time of two hours before sampling. Once excised, all samples were
placed in sealable bags and then an ice-chest to prevent desiccation. Following
collection ", was determined on samples by replicate in the pressure chamber. The
procedure took approximately one hour to complete.
In both years, "'Stem rather than "'Leaf was expected to provide a more robust
indicator of plant water status. For the purpose of examining the effect of irrigation
treatment, "'Stem measurements were made on selected treatments at 0700 HR and 1 300
HR. Four irrigation treatments were selected in 2001 to encompass the range of water
stresses imposed (60 d cutoff, 60 d cutback to 25 % of ET 0, 40 d cutoff, and 20 d
control). In 2002, all treatments were monitored.
1 0 1
4. 1 .4.3 Soil Water Status
A neutron hydroprobe (Model 503, CPN Corp., Pacheco, CA) was used to
evaluate soil volumetric water content (Bv) at depths of 1 5, 30, 60, 90 and 1 20 cm.
Access-tubes were offset 1 5 cm to the side of the drip tape. The probe utilized an
encapsulated 24 1AmlBe source. Soil water measurements were taken weekly starting
either six (2001 ) or seven (2002) weeks preharvest, on a day when irrigation was not
scheduled. Four replications were sampled in 200 1 , five in 2002. Four irrigation
treatments were selected in 2001 to encompass the range of irrigation deficits imposed
(60 d cutoff, 60 d cutback to 25 % of ET 0, 40 d cutoff, and 20 d control). All treatments
were monitored in 2002.
Moisture retention was determined for both soils using the method of Klute
(1 986). Soil was saturated with 0.01 M calcium chloride solution, and then equilibrated
under a range of atmospheric pressure potentials (-20, -30, - 1 00, -500, - 1 500 kPa).
Using the gravimetric water content of the soil at each pressure potential and the
respective bulk density at each site, Bv was estimated; a curve of soil matric potential
(\jfsoi/) against Bv was generated for each site. Available soil moisture (ASM) was
calculated by first determining the maximum amount of plant available water the soil
could supply; this reflected the difference in Bv between field capacity (-20 kPa) and
permanent wilting point (-1 500 kPa). The actual available water content was then
calculated by subtracting measured Bv on any date from that at the permanent wilting
point; this was then expressed as a percent of the maximum available water content in
the soil.
4. 1 .4.4 Yield and Quality Measurements
A single destructive fruit harvest was taken on either August 2 1 st (2001 ) or
August 28th (2002), when control plots had reached approximately 95 % ripe fruit (the
commercial norm). All fruit within a 5 m representative section of row were hand
harvested, and weighed to determine the total yield of each treatment. A sub-sample
was collected and fruit were graded into marketable, green, cull (sunburn and limited
use) and pathological rot categories. Fifty intact marketable fruit per plot were weighed
to calculate mean fruit mass.
1 02
A 5 kg sub-sample of marketable fruit per plot was mechanically juiced and
deaerated under a vacuum of 0.09 MPa to assess quality. Fruit SSC was determined by
temperature-compensated refractometer (RPM-80, Bellingham and Stanley Ltd,
Lawrenceville, GA). Blended colour was quantified by spectrophotometer, and
represented the ratio of green (566 nm) to red (650 nm) . light reflected from the
homogenized juice sample (Agtron E-5, Magnuson Engineers, Inc. , San Jose, CA).
Fruit pH was determined using a pH meter (SS-3, Beckman Coulter, Inc., Fullerton,
CA).
4.1 .5 Statistical Analysis
Experimental data for each season were subjected independently to ANOV A
tests using the SAS GLM procedure. Separation of means was made on significant
ANOV A tests using the LSD method (p < 0.05). Arc-sine transformations were made
where appropriate. Orthogonal contrasts were used to compare the means of specific
irrigation treatments primarily against the control . SAS regression procedures tested the
significance of linear and non-linear models in describing the response of fruit yield and
quality to the percent of ETo applied during ripening. Correlation analysis was used to
report linear associations between variables of interest.
4. 2 2001 Results
The amount of water applied in the final 60 days preharvest ranged from 0 mm
for the severe 60 d cutoff to 3 1 4 mm for the 20 d cutoff (Table 4.4). This was
equivalent to 0 and 82 % of ET 0 during this period, respectively. Based on the
assumption that late-season water loss from processing tomatoes is only 80-90 % of ETo
(Phene et al. , 1 986), all treatments represented some degree of deficit irrigation strategy
during fruit ripening. The earliest ripe fruit were observed approximately one week
after the 60 d treatments were initiated.
1 03
Table 4.4. Irrigation treatment schedule during the final 60 days preharvest.
Treatment schedule Z
Full cutoff implemented 20 days preharvest (control)
Full cutoff implemented 30 days preharvest Full cutoff implemented 40 days preharvest Full cutoff implemented 60 days preharvest (severe stress)
Cutback to 50 % of ET 0 implemented 40 days preharvest Cutback to 25 % of ET 0 implemented 40 days preharvest Cutback to 25 % of ETo implemented 60 days preharvest
Z all cutback treatments had full cutoff at 20 days
4.2.1 Plant Canopy Cover Measurements
Treatment
applied
water (mm)
3 1 4 2 1 8 1 43
0 229 1 86 79
% of ETo
applied in last 60 days
82 57 37 0
60 48 20
Plant canopy measured using the camera technique never reached 1 00 %,
because although adjacent rows were observed to touch, differences in leaf arrangement
resulted in a small portion of bare ground showing in every instance; the computer
software program associated with this technique calculated percent canopy as the
proportion of photosynthetically-active tissue to that which was not (bare ground). On
no date was percent canopy cover significantly affected by irrigation treatment (Table
4.5); the average difference in canopy cover between the control and severe stress plots
(representative of the boundary treatments) was less than 2 %. Across treatments,
canopy cover declined 1 5 % seasonally, most noticeably close to harvest.
Table 4.5. Effect of irrigation treatment on plant canopy cover.
Irrigation treatment Plant canopy cover (%) by days preharvest
20 d cutoff 40 d cutoff 60 d cutoff 25% at 60 d
SE p -value
ns non-significant at p < 0. 1 0
49
70 70 70 67
ns
42 35
67 67 70 66 70 68 68 66
2 2 ns ns
1 04
28 21 14
62 62 62 61 60 62 61 60 62 61 60 63
2 2 ns ns ns
7
57 54 54 52
2 ns
4.2.2 Plant Water Status
4.2.2 . 1 Effect of Sampling Technique
At 0700 HR, 'l'Stem readings in control plots were generally greater (less negative)
than those for 'l'Lea/ (Fig. 4.2). Seasonally, this difference averaged approximately 0.20
MPa, and was highly significant with the exception of measurements made on two of
the eight occasions. On the first of these occasions (35 days preharvest), the minimum
overnight air temperature was lower than normal for this experiment (lower
temperatures would reduce the impact of the bagging technique by decreasing plant
transpiration during the early morning). On the second occasion ( 1 4 days preharvest),
'l'Lea/ was measured at 0500 HR, by design earlier than on the other sampling occasions
and consistent with conventional predawn determination of plant water status. As
expected, on this date the two sampling techniques resulted in similar '1'.
-0. 1 ....-------------.,
-0.6 I -0.7 ....L..,..---..---r---r--,--r-, --r-, --,....J
49 42 35 28 21 1 4 7 o
Days preharvest
Fig. 4.2. Effect of sampling technique on plant water potential ("') in control plots at 0700 HR. Technique means are given ± standard error. Sampling techniques were "'Stem ('Y-) and "'Leaf (---0---) . "'Leaf at 14 days preharvest was measured at 0500 HR (predawn).
As with early morning sampling, 'l'Stem readings taken at 1 300 HR (approximately
solar noon) were consistently greater than those taken simultaneously for 'l'Lea/- This
1 05
trend was observed in both boundary treatments (Fig. 4.3a, b). The greatest difference
between \j!Stem and \j!Leafwas 0.67 MPa, recorded for the severe stress treatment 1 4 days
preharvest. The largest sampling variability was associated with measuring \j!Leaf.
-0.4
-0,5
/-I���� -0,6 ---Ct:l a.. -0.7 �
(b) Severe Stress
Ct:l -0.8 I···· n: .. . . � :;::; c: -0.9 Q)
:2:---_�, +-' 0 ¥ a. - 1 .0 "-Q) +-' '0:: Ct:l - 1 . 1 3: -� -c: -1 .2 Ct:l a..
-1 .3
-1 .4
-1 .5 49 42 35 28 21 1 4 7 0 49 42 35 28 21 1 4 7 0
Days preharvest
Fig. 4.3. Effect of sampling technique on plant water potential ('V) i n control plots (a) and severe stress plots (b) at 1 300 HR. Technique means are given ± standard error. Sampling techniques were 'VStem (-.. -) and 'VLeaf (---0---).
4.2.2.2 Effect of Irrigation Treatment
Stem water potential was adopted as the more reliable indicator of plant water
status as affected by irrigation treatment. Treatments not already implemented were
assumed to have \j!Stem equivalent to that of the control. Decreasing water application
reduced 0700 HR \j!Stem during fruit ripening. This effect was generally significant with
the exception of measurements taken 35 and 1 day preharvest (Table 4.6). As
previously noted, a lower than normal overnight air temperature was reported on the
first of these dates. Highest \j!Stem was reported for control plots on all sampling dates;
the 60 d cutoff treatment resulted in the lowest (most negative) \j!Stem' Similarly, decreasing water application significantly reduced 1 300 HR '!'Stem
during fruit ripening. Again, the control treatment had consistently higher '!'Stem' At
harvest, the difference between the control and most severe cutoff treatment was 0.20
MPa. All treatments showed some degree of seasonal decline when measured at either
1 06
0700 HR or 1 300 HR. Across the final six weeks preharvest, average "'Stem at 0700 HR was -0.25, -0.3 1 , -0.37 and -0.33 MPa and at 1 300 HR -0.57, -0.64, -0.75 and -0.68 MPa
in the 20 d (control), 40 d and 60 d cutoff, and the 60 d cutback, both respectively.
Table 4.6. Effect of i rrigation treatment on "'Stem measured at 0700 HR and 1 300 HR.
Irrigation Stem water potential (MPa) by days preharvest treatment 49 42 35 28 21 14 7 1
-------------------------------------- at 0700 HR ----------------------------------------
20 d cutoff -0.28 a -0.26 a -0.21 -0.29 a -0.23 a -0.22 a -0.24 a -0.34 40 d cutoff -0.30 a -0.30 b -0.31 bc -0.36 b -0.39 60 d cutoff -0.39 b -0.43 b -0.29 -0.36 b -0.35 c -0.34 c -0.39 b -0.41 25% at 60 d -0.29 a -0.34 ab -0.27 -0.34 ab -0.29 b -0.28 ab -0.39 b -0 .41
SE 0 .02 0.02 0.02 0.01 0.01 0.01 0.02 0.01 p -value ns ns
------------------------------------ at 1 300 HR -----------------------------------------
20 d cutoff -0.58 ab -0.46 a -0.48 a -0.54 a -0.61 a -0.56 a -0.63 a -0.66 a 40 d cutoff -0.66 a -0.63 a -0.73 b -0.76 bc -0.79 b 60 d cutoff -0.63 b -0.70 c -0.70 b -0.83 b -0.73 b -0.73 b -0.83 c -0.86 b 25% at 60 d -0.51 a -0.56 b -0.63 b -0.56 a -0.70 b -0.75 b -0 .71 b -0.79 b
SE 0 .02 0.03 0 .04 0.03 0.01 0.02 0.02 0.02 p -value ** ***
ns , * , ** , *** non-significant at p < 0 . 10 , or significant at p < 0.1 0 , 0.05 or 0 .01 , respectively; means within columns separated using the Least Significant Difference (LSD) test, p < 0 .05
4.2.3 Soi l Water Status
Across treatments, highest ()v was generally observed at a depth of 30 cm; drip
tape was buried at 20 cm. Although processing tomatoes are capable of very deep
rooting, the majority of active water-gathering roots are located within the top 50 cm of
soil (Rendon-Poblete, 1 980). Volumetric water content was therefore averaged across
the 1 5, 30 and 60 cm monitoring depths to provide an integrated indication of soil water
status in the main zone of root activity. The two remaining depths (90 and 1 20 cm)
were also averaged for comparative purposes. There was no evidence of a water table
in the top 1 20 cm. Pressure plate analysis showed that ()y at field capacity was
equivalent to 35 .2 % and at permanent wilting point 1 3 .5 %.
The 60 d treatments were implemented approximately two weeks prior to the
first measurement of ()y; early-season dry down in these plots was therefore not well
1 07
characterized. However, water application still significantly affected Bv in the top 60
cm during the remaining six weeks. Volumetric water content was consistently lowest
in the 60 d cutoff and cutback (Fig. 4 .4a). Both of the 60 d treatments lost a steady
amount of soil water between six and four weeks preharvest, with a relatively slow rate
of decline across the final four weeks. On all dates, the two 60 d treatments had
significantly lower Bv than control plots. Following the implementation of the 40 d
cutoff, soil water availability in these plots declined rapidly, resulting in significantly
lower Bv than the control during the final 28 days preharvest. Volumetric water content
did not decline at a substantial rate in the control treatment until cutoff 20 days
preharvest, suggesting that irrigation was approximately equal to plant use during this
period (as expected).
26 �------------------------�
24
...-.. -;§2. e.... 22 ...... c 2 c 20 8 .... 2 1 8 co � t> :s 1 6 (]) E :::J o 1 4 >
1 2
(a) 0-60 cm (b) 60-120 cm
42 35 28 21 1 4 7 o 42 35 28 2 1 1 4 7 o
Days preharvest
Fig. 4.4. Effect of irrigation treatment on soil volumetric water content (9v) in the 0-60 cm (a) and 60-120 cm (b) depth ranges. Treatment means are given ± standard error. Irrigation treatments were full cutoff at 60 days (.), 40 days ( A ), 20 days ( T , control), and cutback to 25 % ET 0 at 60 days (*) . 9v at field capacity was at 35.2 % and at the permanent wilting point was 13.5 %.
When averaged across the final six weeks of fruit ripening, Bv in the top 60 cm
was approximately 22, 20, 1 7 and 1 7 % in the 20 d (control), 40 d and 60 d cutoff, and
the 60 d cutback, respectively. In most instances, By in the top 60 cm was correlated
with '!'Stem (0700 HR and 1 300 HR) during the final three weeks (after all treatments had
been initiated); higher Bv resulted in higher (less negative) '!'Stem. Significant correlation
1 08
coefficients ranged from between 0.58 to 0.72 at 0700 HR, and 0.58 to 0.80 at 1 300 HR; there was no trend over time.
Measurements were not made in the 60- 1 20 cm depth range until the second
week of monitoring (35 days preharvest). On remaining sampling occasions, soil
drying patterns were similar to the top 60 cm, although 8v was consistently less through
the 60- 1 20 cm depth range (Fig. 4.4b). A number of late season 8v readings for the 60
d treatments were below the predicted permanent wilting point; however, as noted the
majority of root activity was not expected at these depths. Available soil moisture
(AS M) and soil matric potential Co/Soil) were calculated across the final six weeks of fruit
ripening (Table 4.7). Across this period, the control treatment had an average ASM content of 37 %, equivalent to a '!'Soil of approximately -91 kPa. This represented a
more modest degree of water stress when compared to 1 6 % ASM in the 60 d cutoff,
equivalent to a '!'Soil of approximately -493 kPa.
Table 4.7. Effect of irrigation treatment on available soil moisture (ASM) and soil matric
potential ("'Soil) across the final six weeks preharvest.
Final six week average
Irrigation Treatment ASM (%)
20 d cutoff 37 40 d cutoff 30 60 d cutoff 1 6 25% at 60 d 1 6
4.2.4 Fruit Yield and Qual ity Measurements
4Jsoil (kPa)
-91 - 152 -493 -492
The effect of irrigation treatment on tomato yield and fruit quality is
summarized in Table 4.8 .
1 09
Table 4.8. Effect of i rrigation treatment on tomato yield and fruit quality at harvest.
Irrigation Total Marketable Soluble Brix Blended Fruit
treatment yield yield solids yield colour z pH
(Mg ha-1 ) (Mg ha-1 ) CBrix) (Brix Mg ha-1 )
20 d cutoff 98 ab 91 ab 5.5 c 5.0 25.4 4.35 30 d cutoff 99 a 90 ab 5.5 c 4.9 23.6 4.37 40 d cutoff 88 bc 8 1 bc 5.9 ab 4.7 23.0 4.34 60 d cutoff 80 c 73 c 6 . 1 a 4.4 23.0 4 .36 50% at 40 d 95 ab 87 ab 5.6 bc 4.8 24.6 4.36 25% at 40 d 97 ab 92 a 5.6 bc 5.2 24.2 4.36 25% at 60 d 91 ab 84 abc 5.9 a 4.9 22.4 4 .35
SE 1 .6 1 .7 < 0.1 < 0.1 0.3 0.01 p -value *** ns ns ns
Z dimensionless unit, lower value indicates more intense red colour in fruit ns , **, *** non-significant at p < 0 . 1 0, or significant at p < 0.05 or 0.01 , respectively; means withi n columns separated using the Least Significant Difference (LSD) test, p < 0 .05
4.2.4. 1 Total and Marketable Fruit Yield
Irrigation application during the final 60 days preharvest significantly affected
total fruit yield (Table 4.8). Yield was greatest for the 30 d cutoff treatment, although
was not significantly different to the control treatment. Lowest yield was obtained
under the 60 d cutoff, which compared poorly to the majority of other treatments .
Separate contrasts also confirmed significantly (p < 0. 1 0) lower fruit yield in both the
40 d cutoff and 60 d cutback compared to the control treatment. Treatment differences
in marketable yield were largely consistent with those for total yield; the percent of
yield that was marketable was approximately 93 % for both the control and 60 d cutoff.
Total and marketable yield were strongly correlated (p < 0.0 1 , r = 0.97).
Significant linear regression models were fitted to both total and marketable
yield (p < 0.05, ? = 0.67 and 0.63 , respectively), with similar slopes in response to
percent of ETo applied (Fig. 4.5a, b). Inclusion of quadratic or cubic terms was not
statistically justified. The linear model confirmed that the percent of ETo applied in the
final 60 days preharvest had a greater effect on both yield measures than the application
strategy (cutoff or cutback). For each 20 % reduction in ETo applied during ripening,
fruit yield declined by approximately 3 Mg ha-I . Total yield was positively correlated
to average Bv (top 60 cm) across the final six weeks of ripening (p < 0.0 1 , r = 0.56);
higher soil water availability was generally associated with higher fruit yields.
1 1 0
Similarly, average '!'Stem measured at 0700 HR and 1 300 HR across the same period was
also positively correlated with total yield (p < 0.0 1 , r = 0.50 and 0.55, respectively); less
negative plant water status was generally associated with higher fruit yields.
1 05
1 00
+ 95
'7 co .c Cl 90 lie � .. "0 Q)
85 '>' ::: :::J .... LL
80 - •
75 -
70 0 20 40
(a) Total
• � ,-0
r2 = 0.67
60 80
-
-
-
•
0
lie ' "
20
"
% of ETa applied in last 60 days
+
40
(b) Marketable
" " � . " " /
" ( = 0.63
o
60 80
Fig. 4.5. Effect of irrigation treatment on total (a) and marketable (b) fruit yield. Treatment means are shown. Irrigation treatments were full cutoff at 60 days (.) , 40 days ( ... ), 30 days (.), 20 days ( T , control), and cutback to 25 % of ETa at 60 days ( *), 25 % of ETa at 40 days (+), and 50 % of ETo at 40 days (0). The equation of the fitted linear regression line ( . . . ) for total yield was y = (85.3 ±3.3) + (0.1 703 ±O.0539).x and for marketable yield y = (78.3 ±3.6) + (0.1 776 ±O.0614).x.
4.2.4.2 Fruit Soluble Solids
Irrigation application during the final 60 days preharvest significantly affected
SSC of marketable fruit (Table 4.8). Fruit SSC was highest for the 60 d cutoff
treatment, although was not significantly different to the 60 cutback or 40 d cutoff.
Lowest fruit SSC was obtained with the 20 d and 30 d cutoffs. Fruit SSC was
negatively correlated to total yield (p < 0.0 1 , r = -0.66); for every 0. 1 aBrix increase in
fruit SSC there was a decline in total yield of 2 Mg ha-I .
A significant linear regression model was fitted for SSC against percent of ET 0
applied (p < 0.02, � = 0.73); application amount appeared more important than
application strategy (Fig. 4.6). The inclusion of quadratic or cubic terms was not
statistically justified. For each 20 % reduction in ETo applied during fruit ripening, fruit
1 1 1
SSC increased by approximately 0. 1 5 °Brix. Fruit SSC was negatively correlated to
average By (top 60 cm) across the final six weeks of ripening (p < 0.0 1 , r = -0.64);
lower soil water availability was generally associated with higher SSc. Similarly,
average "'Stem measured at 0700 HR and 1 300 HR across the same period was also
negatively correlated with fruit SSC (p < 0.0 1 , r = -0.54 and -0.68, respectively); more
negative plant water status was generally associated with higher SSC.
6.1 ... .
6.0
'. * 5.9 ..-... x ·c
0) 5.8 ° -() (j') (j') 5.7 -:'= ::J .... u. 5.6 + �,, _ . r2 = O.73
5.5 •
5.4
0 20 40 60 80
% of ETo applied in last 60 days
Fig. 4.6. Effect of irrigation treatment on fruit soluble solids concentration (SSC). Treatment means are shown. Irrigation treatments were ful l cutoff at 60 days (.) , 40 days (A ), 30 days (.), 20 days ( T , control), and cutback to 25 % of ET 0 at 60 days (*), 25 % of ETo at 40 days (+), and 50 % of ETo at 40 days (0). The equation of the fitted linear regression line ( . . . ) was y = (6.06 ±O.1 1 ) - (0.0077 ±O.0021 ).x.
4.2.4.3 Brix Yield
Brix yield was not significantly affected by irrigation application, averaging
4.88 Mg Brix solids ha'l across treatments (Table 4.8). This was largely consistent with
the contrasting effects of water stress on the two parameters comprising Brix yield;
increasing water stress reduced marketable yield but increased fruit SSc. However, this
'dilution-effect' relationship did not appear unlimited; the contrast comparing only the
border treatments was significant (p < 0.06), confirming comparatively lower Brix
yields under severe water deficit. Brix yield was more strongly correlated with
marketable yield than with SSC (p < 0.05, r = 0.90 and -0. 1 6, respectively).
1 1 2
The response of Brix yield to percent of ETa applied was not well characterized
using linear, quadratic or cubic regression. A linear plateau model appeared to provide
the most logical fit of the data, although was itself limited because a large number of
data points fell outside of the range at which Brix yield began to plateau. Subsequently,
not all of the terms in the model could be estimated simultaneously. However, it
appeared that the 'threshold' value at which Brix yield was sacrificed (i.e. where
marketable yield decline was disproportionate to SSC increase) occurred between the
first and second treatment levels (0-20 % of ETa applied). Using a conservative
estimate of the percent of ETa at which the plateau began (20 %), a model was
developed for illustrative purposes only (Fig. 4.7).
5.2
5.0
4.8
"0 Q) ">, 4.6 x ·c
co
4.4 •
I o
+
,JIE . - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - - - - - - - - -.- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
20 40 60 80
% of ETa applied in last 60 days
Fig. 4.7. Effect of irrigation treatment on Brix yield. Treatment means are shown. Irrigation treatments were full cutoff at 60 days (e), 40 days ( .& ), 30 days (_), 20 days ( T , control), and cutback to 25 % of ETa at 60 days (*), 25 % of ETa at 40 days (+), and 50 % of ETa at 40 days (0).
4.2.4.4 Blended Fruit Colour
Blended fruit colour was not significantly affected by irrigation application,
although fruit grown under higher stress conditions appeared to have numerically lower
colour scores (Table 4.8); a lower score was consistent with a higher red colour
intensity. Fruit colour was not correlated to the yield or percent of green fruit observed
1 1 3
(consistent with harvesting at commercial maturity). A significant linear regression
model was fitted for blended fruit colour against percent of ETo applied (p < 0.06, r2 =
0.62); application amount appeared more important than application strategy (Fig. 4.8).
Inclusion of quadratic or cubic terms was not statistically justified. Despite these
observations, all treatments achieved colour scores that were considered acceptable for
processing (scores < 39).
26�------------------------------------------------�
.... :::J o
25
o 24 <.>
"0
� 23 c a.> co
22 -
<>
+ r2 = 0 .62
•
•
*
21��----�----.-----r----.----�----�----�----.-� o 20 40 60 80
% of ET 0 applied in last 60 days
Fig. 4.8. Effect of irrigation treatment on blended fruit colour. Treatment means are shown. Irrigation treatments were full cutoff at 60 days (e), 40 days ( .. ), 30 days (_), 20 days ( T , control), and cutback to 25 % of ETo at 60 days (*), 25 % of ETo at 40 days (+), and 50 % of ETo at 40 days (0). The equation of the fitted linear regression
line ( . . . ) was y = (22.4 ±O.6) + (0.0289 ±O.01 14).x.
4.2.4.5 Fruit pH
Fruit pH was not significantly affected by irrigation treatments imposed during
the final 60 days preharvest (Table 4.8). Across treatments, fruit pH averaged 4.36.
4.2.4.6 Mean Marketable Fruit Mass and Fruit Number
The effect of irrigation treatment on the mean mass of marketable fruit (MFf) was not significant. However, a consistent trend resulting in declining MFf with reduced
water application was observed for all cutoff treatments. Contrasts between the control
1 1 4
and 60 and 40 d cutoffs were both significant (p < 0.03 and 0.06, respectively), while
there was no difference between the control and 30 d cutoff. Marketable yield was
positively correlated with MFf (p < 0.02, r = 0.42); greater fruit mass generally resulted
in higher yield outcomes. A significant linear regression model was fitted for MFf against percent of ETo applied (p < 0.02, r2 = 0.70); application amount appeared more
important than application strategy (Fig. 4.9). The inclusion of quadratic or cubic terms
was not statistically justified.
80
78 -
76
74 "-'3 � 72 Ol '-" "::£� 70
68
• 66 -
64 o
+--
20 40
o . - -- --- - r2 = 0.70
60
% of ETo applied in last 60 days
80
Fig. 4.9. Effect of irrigation treatment on mean marketable fruit mass (MF,). Treatment means are shown. Irrigation treatments were full cutoff at 60 days ( .), 40 days ( . ), 30 days (_), 20 days C T , control), and cutback to 25 % of ETo at 60 days C*), 25 % of ETo at 40 days (+), and 50 % of ETo at 40 days (0). The equation of the fitted linear
regression l ine ( ... ) was y = (68.6 ±1 .7) + (0.1058 ±O.0307).x.
The number of marketable fruit per plant (as estimated from marketable yield
and MFf) was not significantly affected by the irrigation treatments imposed (data not
shown). Although the 60 d cutoff had numerically lower fruit number than the control,
the contrast between these border treatments was not significant.
1 1 5
4.2.4.7 Yield Components
Yield components were adjusted to a percent of total fruit yield to allow a
standardized comparison of crop maturity unaffected by treatment-related differences in
productivity (Fig. 4. 1 0) .
-� 0 -Cf) 90 ..... C Q) C 0 a. E 0 () 32 80 Q) >=
I rrigation treatment
Fig. 4.10. Effect of irrigation treatment on yield components at maturity. Yield components were marketable (_), rots ( ) , culls (0) and greens (181).
Irrigation treatments did not significantly affect the percent of fruit in any yield
category (marketable, rot, cull and green). Across treatments, marketable yield
averaged approximately 93 % of total yield. The predominant rot species was water
mold. The contrast comparing only the percent rots between border treatments was
significant (p < 0.06); rot incidence was approximately 2 % greater under severe water
deficit. However, rots remained a small constituent of total yield, averaging less than 3
% across treatments. Percent culls averaged less than 2.5 % of total yield. Across
treatments, sunburn-affected fruit accounted for approximately 75 % of total culls; the
remainder was limited-use fruit (fruit that were over ripe). Separately, there was no
effect of irrigation treatment on the incidence of either sunburn or limited-use fruit.
There was also no correlation between percent canopy cover during ripening and
sunburn incidence on fruit at maturity. Green fruit comprised less than 3 % of total
yield; this level was comparable to that in commercial fields at maturity. Large
1 1 6
sampling variability within treatments was observed for the rot, cull and green
components (cv = 50, 8 1 and 74 %, respectively).
4.3 2002 Results
The amount of water applied in the final 50 days preharvest ranged from 0 mm
for the severe 50 d cutoff to 253 mm for the 20 d cutoff (Table 4.9). This was
equivalent to 0 and 77 % of ETo during this period, respectively. All treatments
therefore represented some degree of deficit irrigation strategy during fruit ripening
(Phene et al. , 1 986). In the 1 0 days before treatments were implemented (50-60 days
preharvest) 7 1 mm of water was applied to all plots. First ripe fruit were observed
concurrently with the implementation of the early stress treatments.
Table 4.9. I rrigation treatment schedule during the final 50 days preharvest.
Treatment schedule Z
Full cutoff implemented 20 days preharvest (control)
Full cutoff implemented 30 days preharvest Full cutoff implemented 50 days preharvest (severe stress)
Cutback to 25 % of ETa implemented 50 days preharvest
Z cutback treatment had full cutoff at 20 days
4.3 . 1 Plant Water Status
4.3. 1 . 1 Effect of Sampling Technique
Treatment
applied water (mm)
253 1 72 0
63
0/0 of ETo
applied in last 50 days
77 52
0 1 9
Unlike the previous season, bagging leaves overnight in control plots for 0700
HR sampling did not consistently increase "'Stem compared to "'Leaf (Fig. 4. 1 1 ).
Seasonally, the difference between the two sampling techniques varied from 0.02 to
0. 1 8 MPa, although measurements made on four of the seven sampling occasions were
not significantly different. On the first of these occasions (27 days preharvest) "'Leaf
was measured at 0500 HR, consistent with conventional predawn determination of "'.
No difference was therefore expected on this date. On the second occasion (20 days
preharvest), the air vapour pressure and relative humidity were lower than normal for
1 1 7
this experiment (lower pressure and humidity would reduce the impact of the bagging
technique by decreasing plant transpiration). On the third and fourth occasions ( 1 3 and
7 days preharvest) overnight and early morning weather observations were within the
normal range expected. Leaf water potential readings were not taken 4 1 days
preharvest.
-0.2 ..,-------------,
-0.3 .-to c.. :::2!
-0.4 -
co <:; c Q) -0 a.
-0.5 1:
.... Q) -
l to � -0.6 -c to c..
'2: -0 .7
-0.8 .... ---.--,-.,--,---,-----,r-' 42 35 28 21 14 7 0
Days preharvest
Fig. 4.1 1 . Effect of sampling technique on plant water potential ('!') in control plots at 0700 HR. Technique means are given ± standard error. Sampling techniques were
'!'Stem (- .. -) and \jILeaf (---0---). '!'Leaf at 27 days preharvest was measured at 0500 HR (predawn).
When measured at 1 300 HR, 'VStem was consistently higher than 'VLeq(, this trend
was observed in both boundary treatments (Fig. 4. 1 2a, b). The greatest difference
between 'VStem and \jILea/ was 0.55 MPa, recorded for the control treatment 41 days
preharvest. Surprisingly, variability associated with sampling '!ILea/ WaS often similar to
1 1 8
-co a. � ro � c Q) ..... 0 0.. � Q) ..... co 3: ..... c co
a.
-0.4
-0.5
-0.6
-0.7
-0.8
-0.9
- 1 .0
-1 . 1
- 1 .2
-1 .3
- 1 .4
(a) Control
Y\ lyI----r
" "
1
�, � / \ 1 I--_ 1:[
" ,
� �
-1 .5 -'-r---r---r----r---.---.---r-' 42 35 28 21 14 7 o
I--- - -� -J:{:
(b) Severe Stress
;1\ !�
''2:" l----�, .. ,, " "
'I
42 35 2 8 21 1 4 7 o
Days preharvest
Fig. 4.1 2. Effect of sampling technique on plant water potential ("') in control plots (a) and severe stress plots (b) at 1300 HR. Technique means are given ± standard error.
Sampling technique were "'Stem (-T -) and "'Leaf (---0---).
4.3. 1 .2 Effect of Irrigation Treatment
In following the 2001 practice, 'VStem was adopted as the more reliable indicator
of plant water status as affected by irrigation treatment. Decreasing water application
reduced 0700 HR 'VStem during fruit ripening. This effect was generally significant, with
the exception of measurements taken 4 1 and 7 days preharvest (Table 4 . 1 0). Overnight
and early morning weather observations on these two sampling dates were within the
normal range observed for this experiment. Highest 'VStem was generally reported for
control plots; the 50 d cutofftreatment consistently resulted in the lowest 'VStem.
Similarly, decreasing water application reduced 1 300 HR 'VStem during fruit
ripening. As with morning sampling, this effect was generally significant, with the
exception of measurements taken 4 1 and 1 3 days preharvest. On these occasions, mid
morning and early afternoon weather observations were within the normal range
observed for this experiment. The control treatment had the highest 'VStem, although this
was not different to 30 d cutoff until late in the season. At harvest, the difference
between the control and most severe cutoff treatment was 0. 1 8 MPa. All treatments
showed some degree of seasonal decline when measured at either 0700 HR or 1 300 HR.
1 1 9
Across the final six weeks preharvest, average '!'Stem at 0700 HR was -0.4 1 , -0.44, -0.53
and -0.47 MPa and at 1 300 HR -0.74, -0.77, -0.92 and -0.88 MPa in the 20 d (control),
30 d and 50 d cutoff, and the 50 d cutback, both respectively.
Table 4.1 0. Effect of irrigation treatment on '!'Stem measured at 0700 HR and 1300 HR.
Irrigation Stem water potential (MPa) by days preharvest
treatment 41 34 27 20 13 7 1 ---------------------------------- at 0700 HR -----------------------------------
20 d cutoff -0.28 -0.26 a -0.43 a -0.51 a -0.36 a -0.54 -0 .51 a 30 d cutoff -0.31 -0.25 a -0.40 a -0.55 a -0.43 ab -0.57 -0.59 b 50 d cutoff -0.35 -0.38 b -0.59 b -0.62 b -0.47 b -0.65 -0.66 c 25% at 50 d -0.33 -0.34 b -0.49 a -0.55 a -0.39 ab -0.58 -0.58 b
SE 0.01 0.01 0.02 0.01 0 .02 0.02 0.02 p -value ns *h ** ns
--------------------------------- at 1 300 HR -----------------------------------20 d cutoff -0.64 -0.58 a -0.74 a -0.81 a -0.75 -0.78 a -0.89 a 30 d cutoff -0.61 -0.60 a -0.68 a -0.86 ab -0.78 -0.90 ab -0.95 ab 50 d cutoff -0.72 -0.76 a -0.90 b -1 .02 b -0.88 -1 .06 c -1 .07 e 25% at 50 d -0.68 -0.64 b -0.88 b -0.99 b -0.82 -0.98 bc -1 .02 be
SE 0.02 0.02 0.03 0.03 0.02 0.03 0.02 p -value ns ** *** ns ***
ns , *, **, *** non-significant at p < 0 . 10 , or significant at p < 0.1 0, 0 .05 or 0 .01 , respectively; means within columns separated using the Least Significant Difference (LSD) test, p < 0 .05
4.3 .2 Soil Water Status
Across treatments, highest By was generally observed at a depth of 60 cm; drip
tape was buried at 20 cm. In following the 2001 practice, By was averaged across the
1 5 , 30 and 60 cm depths to provide an integrated indication of soil water status in the
main zone of root activity, and also across the two remaining depths (90 and 1 20 cm).
There was no evidence of a water table in the top 1 20 cm. Pressure plate analysis
showed that By at field capacity was equivalent to 39.0 % and at permanent wilting
point 1 6.7 %.
A calibration error in the standard count data (integral to calculating (}y) was
found on the first three sampling occasions (48, 4 1 and 34 days preharvest); By determinations on these dates were subsequently omitted. However, unadjusted neutron
counts during this initial three week period remained consistent with trends established
1 20
in 200 1 ; following the implementation of the 50 d treatments, there was a rapid initial
decline in neutron count in the top 60 cm depth (lower neutron count is consistent with
comparatively less soil water on a given date). As expected, 20 d and 30 d cutoff
treatments had similar neutron counts through this initial three week period. When the
inconsistency with standard counts was corrected, water application during the
remaining four weeks preharvest still significantly affected Bv in the top 60 cm. During
this period, (}v was consistently lowest in the 50 d cutoff (Fig. 4. 1 3a). On all dates, the
50 cutoff had lower significantly (}y than control plots. The 50 d cutback treatment also
had significantly lower (}y than the control plots, with the exception of the final
sampling 1 day preharvest. Volumetric water content did not decline at a substantial
rate in the control treatment until cutoff 20 days preharvest; as in 2001 , this suggested
that irrigation was approximately equal to plant use during this period.
Fig. 4.1 3. Effect of irrigation treatment on soil volumetric water content (8v) in the 0-60
cm (a) and 60-120 cm (b) depth ranges. Treatment means are given ± standard error. Irrigation treatments were full cutoff at 50 days (_), 30 days (e), 20 days ( . , control), and cutback to 25 % ET 0 at 50 days (*). 8v at field capacity was at 39.0 % and at the permanent wilting point was 1 6.7 %.
When averaged across the final four weeks of fruit ripening, (}y in the top 60 cm
was approximately 24, 2 1 , 1 8 and 20 % in the 20 d (control), 30 d and 50 d cutoff, and
the 50 d cutback, respectively. Volumetric water content in the top 60 cm was
consistently correlated with "'Stem (0700 HR and 1 300 HR) during the final three weeks
1 2 1
(after all treatments had been initiated); higher Bv resulted in higher (less negative)
\jIStem. Significant correlation coefficients ranged from between 0.40 to 0.7 1 at 0700 HR, and 0.46 to 0.75 at 1 300 HR; there was no trend over time.
Both the 20 d and 30 d cutoff had slightly higher Bv in the 60- 1 20 cm depth
range than the top 60 cm; drying patterns were however generally similar (Fig 4. 1 3b).
With the exception of the 50 d cutoff, all other treatments had Bv higher than the
predicted permanent wilting point. Little root activity was expected at the lower soil
depths.
As in 200 1 , calculations of ASM and \jIsoil were made (Table 4. 1 1 ) ; however,
because of the calibration error during early monitoring, both variables could only be
estimated across the final four weeks of fruit ripening rather than six. Across this
truncated period, the control treatment had an average ASM content of 3 1 %, equivalent
to a \jIsoil of approximately -248 kPa. This represented a considerably more modest
degree of water stress when compared to 7 % ASM in the 50 d cutoff, equivalent to a
\jIsoil of approximately - 1 096 kPa. The absence of reliable Bv measures during the six to
four week monitoring period clearly misrepresented water availability in both the
control and severe stress treatments during fruit ripening.
Table 4.1 1 . Effect of irrigation treatment on available soil moisture (ASM) and soil matric
potential ("'Soil) across the final four weeks preharvest.
Final four week average
Irrigation Treatment
20 d cutoff 30 d cutoff 50 d cutoff 25% at 50 d
ASM (%)
31 20 7 1 4
4.3.3 Fruit Yield and Quality Measurements
IV soil (kPa)
-248 -51 4
-1 096 -724
The effect of irrigation treatment on tomato yield and fruit quality IS
summarized in Table 4. 1 2.
1 22
Table 4.1 2. Effect of irrigation treatment on tomato yield and fruit quality at harvest.
Irrigation Total Marketable Soluble Brix Blended Fruit
treatment yield yield solids yield colour Z pH
(Mg ha-1 ) (Mg ha-1 ) CBrix) (Brix Mg ha-1 )
20 d cutoff 1 25 a 1 09 a 5.3 b 5.7 25.5 a 4.36 30 d cutoff 1 1 9 a 1 03 a 5.3 b 5.5 25.3 a 4.34 50 d cutoff 1 03 b 88 b 5.9 a 5.2 23.0 b 4.34 25% at 50 d 1 1 7 a 1 06 a 5.5 b 5.8 24.2 ab 4.30
SE 2.3 2.6 < 0.1 0.1 0.4 0.01 p -value *** *** ns ** ns
Z dimensionless unit, lower value indicates more intense red colour in fruit ns , ** , *** non-significant at p < 0 . 10 , or significant at p < 0.05 or 0.01 , respectively; means within columns separated using the Least Significant Difference (LSD) test, p < 0.05
4.3 .3 . 1 Total and Marketable Fruit Yield
Irrigation application during the final 50 days preharvest significantly affected
total fruit yield (Table 4. 1 2). Yield was greatest for the control treatment, although was
not significantly different to 30 d cutoff or 50 d cutback. Lowest yield was obtained
under the 50 d cutoff, which compared poorly to all other treatments. Treatment
differences in marketable yield were largely consistent to those for total yield; the
percent of yield that was marketable was approximately 86 % for both the control and
50 d cutoff treatments. Total and marketable yield were strongly correlated (p < 0.0 1 , r
= 0.93).
A significant linear regression model was fitted to total yield (p < 0.03 , r2 =
0.92) but not marketable yield in response to the percent of ET 0 applied (Fig. 4. 1 4a, b).
Inclusion of quadratic or cubic terms was not statistically justified. The linear model
confirmed that the percent of ETo applied in the final 50 days preharvest had a greater
effect on total yield than the application strategy. For each 20 % reduction in the ETo
applied during ripening, fruit yield declined by approximately 6 Mg ha-! . Total yield
was positively correlated to average Bv (top 60 cm) across the final four weeks of
ripening (p < 0.0 1 , r = 0.62); higher soil water availability was generally associated with
higher fruit yields. Similarly, average '!'Stem measured at 0700 HR and 1 3 00 HR across the
final six weeks was also positively correlated with total yield (p < 0.0 1 , r = 0.82 and
0.8 1 , respectively); less negative plant water status was generally associated with higher
fruit yields.
1 23
1 30 (a) Total (b) Marketable
1 25 - T
1 20 .- � = 0.92 lIE
'7 1 1 5 co ..c -0> :2 1 1 0 -"C CD 1 05 .::;' it == • � .... 1 00 LL
95 -
90 •
85 I 0 20 40 60 80 0 20 40 60 80
% of ETo applied in last 50 days
Fig. 4.1 4. Effect of irrigation treatment on total (a) and marketable (b) fruit yield. Treatment means are shown. Irrigation treatments were full cutoff at 50 days ( .) , 30 days (.) , 20 days ( T , control), and cutback to 25 % of ETo at 50 days (*). The
equation of the fitted linear regression line ( ... ) for total yield was y = (1 04.3 ±2.1 ) + (0.2883 ± 0.0024).x.
4.3.3 .2 Fruit Soluble Solids
Irrigation application during the final 50 days preharvest significantly affected
SSC of marketable fruit (Table 4. 1 2). Fruit SSC was highest for the 50 d cutoff
treatment. Lowest fruit SSC was obtained with the 20 d and 30 d cutoffs. Fruit SSC
was negatively correlated to total yield (p < 0.0 1 , r = -0.69); for every 0. 1 °Brix increase
in fruit SSC there was a decline in total yield of approximately 2 Mg ha- I .
A significant linear regression model was fitted for SSC against percent of ETa
applied (p < 0.08, r2 = 0.72); application amount appeared more important than
application strategy (Fig. 4. 1 5). The inclusion of quadratic or cubic terms was not
statistically justified. For each 20 % reduction in ETa applied during fruit ripening, fruit
SSC increased by approximately 0. 1 5 °Brix. Fruit SSC was negatively correlated to
average (}v (top 60 cm) across the final four weeks of ripening (p < 0.04, r = -0.49);
lower water availability was associated with higher SSC. Similarly, average \j!Stem measured at 0700 HR and 1 300 HR across the six week period preceding harvest was also
1 24
negatively correlated with fruit SSC (p < 0 .0 1 , r = -0.75 and -0.82, respectively) ; more
negative plant water status was generally associated with higher SSC.
6.0
5.9 •
5.8
-><
5.7 .t: cc 0 5.6 '-' () Cl) Cl) 5.5 --·S .... LL 5.4
5.3 •
5.2
5.1 0 20 40 60 80
% of ETa applied in last 50 days
Fig. 4. 15. Effect of irrigation treatment on fruit soluble solids concentration (SSe). Treatment means are shown. Irrigation treatments were full cutoff at 50 days ( .), 30 days (_) , 20 days (T, control), and cutback to 25 % of ETo at 50 days <*). The
equation of the fitted linear regression line ( . . . ) was y = (5.75 ±O.1 3) - (0.0071 ±O.0024).x.
4.3.3 .3 Brix Yield
Brix yield was not significantly affected by irrigation application, averaging
5.55 Mg Brix solids ha- 1 across treatments (Table 4. 1 2). Unlike 200 1 , the contrast
comparing only the border treatments was not significant, despite a similar numerical
increase of approximately 0.5 Mg Brix solids ha-1 in the control. Brix yield was
significantly correlated with marketable yield (p < 0.0 1 , r = 0.88) but not fruit SSc.
The response of Brix yield to percent of ET 0 applied was not well characterized using
linear, quadratic or cubic regression. A linear plateau model was successfully fitted to
the data (p < 0.06; Fig. 4. 16). The inflection point at which a further reduction in
percent of ETo applied was predicted to reduce Brix yield was approximately 1 5 %.
1 25
5.8
";' ('Cl .c (J) 5.6 :g CS (J) X ·c -
CO Cl 5.4 -:E '-'
'0 Q) '>, X ·c 5.2 CO
5.0�'-----�-----I.-----r-----TI----�-----'I----��--� o 20 40 60 80 % of ET 0 applied in last 50 days
Fig. 4. 1 6. Effect of irrigation treatment on Brix yield. Treatment means are shown. Irrigation treatments were full cutoff at 50 days (e) , 30 days (_), 20 days ( T , control), and cutback to 25 % of ETa at 50 days (*). The equation of the fitted linear plateau
regression line ( . . . ) was y = (5. 19 :±O.21 ) + (0.0327 :±O.01 64).x where x < 14.9, and y = 5.67 where x � 14.9.
4.3.3.4 Blended Fruit Colour
Blended fruit colour was significantly affected by irrigation application in the
final 50 days preharvest (Table 4. 1 2). Colour scores were lowest (most red) in the 50 d
cutoff and highest in the 20 d and 30 d cutoffs. Fruit colour was not correlated to the
yield or percent of green fruit observed. A significant linear regression model was
fitted for blended colour against percent of ETo applied (p < 0.05, r2 = 0.88); application
amount appeared more important than application strategy (Fig. 4. 1 7). Inclusion of
quadratic or cubic terms was not statistically justified. Despite these observations, all
treatments achieved colour scores that were considered acceptable for processing.
1 26
.... :::J 0 0 u
:!::: :::J .!= u Q)
u c Q)
CO
26
25
24 -
23 •
o
T -
r2 = 0.88
lIE
20 40 60 80
% of ETo appl ied in last 50 days
Fig. 4.17 . Effect of irrigation treatment on blended fruit colour. Treatment means are shown. Irrigation treatments were full cutoff at 50 days (e ), 30 days (_), 20 days ( T , control) , and cutback to 25 % of ETo at 50 days (*). The equation of the fitted linear
regression l ine ( . . . ) was y = (23.2 ±O.3) + (0.0332 ±O.0069).x.
4.3 .3.5 Fruit pH
Fruit pH was not significantly affected by irrigation treatments imposed during
the final 50 days preharvest (Table 4. 1 2). Across treatments, fruit pH averaged 4.34.
4.3 .3 .6 Mean Marketable Fruit Mass and Fruit Number
Mean mass of marketable fruit was significantly (p < 0.07) affected by irrigation
application in the final 50 days preharvest; MFf was greatest in the 20 d and 30 d cutoff
treatments, and lowest in the severe 50 d cutoff. Marketable yield was positively
correlated with MFf (P < 0.02, r = 0.55); greater fruit mass generally resulted in higher
yield outcomes. A significant linear regression model was fitted for MFf against percent
of ETo applied (p < 0.06, r2 = 0.83); application amount appeared more important than
application strategy (Fig. 4. 1 8). The inclusion of quadratic or cubic terms was not
statistically justified. The number of marketable fruit per plant was not significantly
affected by the irrigation treatments imposed (data not shown). However, the 50 d
1 27
cutoff had significantly (p < 0.06) lower fruit number when compared directly against
the control treatment.
78�-------------------------------------------------,
76
• • 74 2 = 0.83 r
'-·5 L.. 72 -..... 0> -�LL
70 lIE
68 •
66�,-----�-----,r-----�----�----�----'-----�----� o 20 40 60 80
% of ETa applied in last 50 days
Fig. 4.18. Effect of irrigation treatment on mean marketable fruit mass (MFf). Treatment means are shown. Irrigation treatments were full cutoff at 50 days (e), 30 days (.) , 20 days ( T , control), and cutback to 25 % of ETo at 50 days (*). The equation of the
fitted linear regression line ( ... ) was y = (68.2 ±1 .3) + (0.1036 ±O.0258).x.
4.3.3.7 Yield Components
In following the 2001 practice, yield components were adjusted to a percent of
total fruit yield to allow a standardized comparison of crop maturity (Fig. 4 . 1 9) .
Irrigation treatments did not significantly affect the percent of fruit that were
marketable, rots or culls. Across treatments, marketable yield averaged approximately
87 % of total yield. The predominant rot species was water mold. The contrast
comparing only the percent rots between border treatments was significant (p < 0.06);
rot incidence was approximately 5 % for the 50 d cutoff and 2 % for the 20 d control .
Percent culls averaged 5 % of total yield. Although not categorized separately,
sunburn-affected fruit accounted for most culls. By comparison, irrigation treatment
significantly affected the percent of fruit that were green (p < 0.02); at harvest, green
fruit comprised approximately 3 % of total yield for the 50 d treatments compared to 7
% for the 20 d or 30 d treatments. Large sampling variability within treatments was
observed for the rot, cull and green components (cv = 80, 49 and 44 %, respectively).
1 28
-"::§? 0 90 -Cl) ........ C Q) C 0 80 0.. E 0 ()
""0 Q)
70 >=
0 � ...... � ...... � ...... � ...... 1W I rrigation treatment
Fig. 4.19. Effect of irrigation treatment on yield components at maturity. Yield Components were marketable (.) , rots ( ) , culls (0) and greens ([gJ).
4.4 Discussion
4.4 .1 Effect of Water Appl ication on Plant Canopy Cover
Poorly-timed irrigation deficits can cause severe canopy die-back in processing
tomatoes, increasing the likelihood of fruit ' sunburn' . Sunburn is an imperfection on
fruit rendering them unsuitable for whole-peeling. Such symptoms appear when fruit
are directly exposed to solar radiation and overheating of tissue occurs (Adegoroye and
Jolliffe, 1 983). Ideally, growers need to establish a full vegetative canopy early in the
season to support flowering and fruit set, and retain as much of this cover until close to
harvesting. This not only allows fruit to be shaded to some extent (in so doing reducing
the risk of sunburn), but also retains leaf assimilate supplies that can be redistributed to
fruit as they develop (ensuring high fruit quality). Some degree of canopy decline is
unavoidable as plants senesce close to harvest; however, late-season irrigation strategies
that cause excessive die-back should be avoided.
In both seasons full plant cover was achieved before irrigation treatments were
imposed; accordingly, there was no water-stress impediment to maximum canopy
development. In no instance did increasing late-season water stress significantly reduce
1 29
measures of percent canopy made in 200 1 . This suggests a high tolerance to water
deficit without canopy decline; even the most severe 60 d irrigation cutoff did not
accelerate plant die-back or vine collapse. Typically, \jILeaj exceeding -0.6 to -0.7 MPa
causes stomatal closure in tomato (Cerda et al. , 1 979; Rudich et al. , 1 98 1 ), which in
turn can reduce the ability of the plant to regulate its internal temperature; excessively
high temperature can disrupt a number of physiological processes and also cause leaf
tissue to die. Although on several sampling occasions severe stress treatments had \jI
readings exceeding this threshold, these were generally only observed under peak solar
load; plants returned to less negative water potentials overnight. Rudich et al. ( 1 98 1 )
suggested that even well-watered field tomatoes can experience daily fluxes in \jI that
exceed -0.6 MPa. Where \jI consistently exceeds critical thresholds canopy decline may
become problematic; on a lighter-textured soil with low moisture retention, severe
irrigation cutoffs imposed as early as 60 days preharvest may have in fact increased
dieback. Cultivar selection can also have a strong effect on the incidence of canopy
decline; cultivars with a large root system may be able to buffer water deficits to a
greater extent, thereby moderating water stress in the plant.
4.4.2 Effect of Water Appl ication on Plant Water Status
4.4.2 . 1 Effect of Sampling Technique
Traditionally, predawn measurements of \jI have been made to quantify plant
water status when it is close to equilibrium with available soil water (Meyer and Green,
1 980). However, the commercial use of this technique has been limited because
measurements on exposed leaves must occur before sunrise (and increasing water
demand). In the current studies, bagging leaves overnight for early morning sampling
the next day extended the period during which estimates of predawn plant water
potential could be reliably made. This would allow growers to monitor a greater
number of fields between 0500 HR (sunrise) and 0700 HR. The fact that 0700 HR \jIStem
(measured on bagged leaves) was generally less negative than 0700 HR \jILeaj (measured
on exposed leaves) confirmed the rapid effect of radiation and air temperature on
transpiration and plant water potential.
1 30
Meyer and Green ( 1 98 1 ) suggested that the bagging technique can also reduce
short-term environmental variability as compared to actively transpiring leaves.
Greatest variability is frequently observed under peak solar radiation, which is also
concurrent to maximum photosynthetic rates and water demand. However, sampling '"
under these conditions provides an important measure of plant water stress unaffected
by the overnight recovery of water turgor. Bagging leaves for at least two hours before
afternoon measurement reduced sampling variability related to transient environmental
factors; across seasons, the difference between "'Lea! and "'Stem averaged approximately
0.35 MPa. By substantially reducing transpiration, the bagging technique allowed the
water potential of leaves to equilibrate with the stem and roots. This provided a more
reliable indicator of overall plant water status as related to fruit yield and quality. The
overall success of the bagging technique was consistent to findings made in other crops
(Fulton et al. , 200 1 ; McCutchan and Shackel, 1 992). This appeared to be the first work
published specifically on processing tomato.
4.4.2.2 Effect of Irrigation Treatment
Plant-based measurements provide the only true measure of actual water stress
at any given time; methods such as the bagging technique that reduce variation may
then have some practical field application for late-season irrigation management. Early
morning and early afternoon measurements of plant water potential on bagged leaves
confirmed that irrigation treatments caused varying levels of water stress. This was
generally reflective of the percent of ET 0 applied; greater irrigation amount resulted in
lower plant water stress.
Average "'Stem during fruit ripening was consistently correlated to fruit yield and
SSC at commercial maturity. Measurements made in the early afternoon typically
resulted in the strongest relationships with yield and quality. This appeared to reflect
the absence of overnight recovery in leaf turgor, as was observed with many early
morning measurements; inasmuch, overnight recovery tended to moderate differences
between treatments. More negative plant water potential (indicative of greater water
stress) was associated with lower yield and higher fruit SSC. Average 1 300 HR "'Stem in
the range of -0.5 to -0.7 MPa during the final six weeks preharvest was sufficient to
achieve fruit SSC levels > 5.0 °Brix, an acceptable commercial target for many
1 3 1
growers. These results suggest that the use of 'VStem to help augment late-season
irrigation management (particularly when measured under peak solar load) may warrant
further research.
4.4.3 Effect of Water Appl ication on Soi l Water Status
The rapid sensitivity of soil water status to irrigation application has resulted in
a number of soil-based monitoring techniques being investigated to track soil moisture
and manage fruit SSC (Cahn et aI. , 2003 , 2004; Calado et al. , 1 990; Hartz, 1 996; Phene,
1 999) . Soil-based moisture measures are not generally influenced by transient weather
factors, a limitation of traditional '!' measurements. Many growers therefore believe
soil measures to be a more reliable predictor on which to manipulate irrigation
decisions.
Measurements of Bv confirmed that irrigation practices achieved varying levels
of water availability within the soil. This was generally reflective of the percent of ET 0
applied; greater irrigation amount resulted in higher availability. In neither year did Bv
in the top 60 cm decline significantly until after each respective treatment was imposed.
This strongly suggests that the calculated irrigation amounts were sufficient to replace
actual evapotranspiration (as intended). Volumetric water content generally declined at
a rapid rate immediately following the implementation of the each treatment, slowing
thereafter as soil tension increased; increasing soil tension reduces the availability of
water to the plant roots, as the soil matrix holds water more tightly (Brady, 1 990). In several treatments, available soil water was substantially depleted by harvest, although
Bv did not generally drop below the permanent wilting point associated with each soil.
Although prolonged periods just above the permanent wilting point would likely
increase fruit SSC, a disproportionate loss in yield may also result. Such severe water
stresses are seldom recommended for commercial practice.
Average Bv during fruit ripening was not as strongly correlated to yield and fruit
SSC as 1 300 HR 'VStem' As previously described, this appeared to be the result of
bagging leaves for several hours before measurement, which overcame many of the
limitations previously related to sampling exposed leaves. Volumetric water content
was further interpreted in terms of available soil moisture and '!'soi!. However, across
seasons no reliable '!'soi! target during fruit ripening was found. In the first season, '!'soi!
1 32
as modest as approximately -90 kPa was sufficient to achieve fruit SSC > 5 .0 °Brix,
while in the second season '!'Soil as high as approximately -250 kPa was necessary. By
comparison, Cahn et at. (2004) reported that a 'l'soil of - 150 kPa across the final 6 weeks
preharvest was sufficient to achieve fruit SSC > 5 .0 °Brix.
Collectively, these observations indicated that monitoring soil water status did
not offer the best approach from which to base late-season irrigation to accurately raise
fruit SSC with minimal yield loss. Because no irrigation system is entirely uniform and
no soil is completely homogenous, representative readings can only be obtained by
monitoring moisture at various depths and locations within each field; this type of setup
may become impractical to manage. The use of soil-based monitoring systems
therefore appears best targeted as a 'backstop' for over- or under-watering based upon
modelled crop water balances. For this purpose, electrical resistance blocks and soil
tensiometers appear most appropriate, because they are affordable, practical, and can be
automated. Despite improvements in the reliability and accuracy of 'I' measures
through bagging, growers are unlikely to use such techniques as an irrigation back-stop;
this is because measuring '!'stem can be time consuming and requires a greater level of
user-expertise.
4.4.4 Effect of Water Appl icatio n on Yield and Qual ity Measurements
4.4.4. 1 Total and Marketable Fruit Yield
Total and marketable fruit yield generally decreased with reduced water
application during ripening. These findings were consistent with previous studies under
both arid and humid climates (Cahn et al. , 200 1 , 2003 ; Calado et al. , 1 990; May and
Gonzales, 1 994, 1 999; May et al. , 1 990; Mitchell et al. , 1 99 1 a, b; Murray, 1 999;
Renquist and Reid, 2001 ; Wudiri and Henderson, 1 985; Zegbe-Dominguez et al. , 2003) .
Most report that increasing water deficit during fruit sizing and ripening reduces yield,
although the magnitude of loss has varied considerably depending on the timing and
severity of deficit and by soil and environmental factors.
In the current studies, means separation or contrasts confirmed that four of the
five treatments applying less than 40 % of ETo in the final 60 days had significantly
lower fruit yields than the control treatment; this included cutoff treatments initiated 40
1 33
days or more preharvest and cutback treatments initiated 60 days preharvest. By
comparison, there was a relatively high tolerance to water deficits imposed within the
final 40 days (approximately 6 weeks) of ripening. In no trial did the application
strategy itself (cutoff or cutback) appear to have an overall statistical influence on fruit
yield (total or marketable). In as much, the percent of ETo applied appeared to be a
useful predictor of yield once fruit had set and begun to ripen.
Lower yield was primarily related to smaller fruit size, although severe irrigation
cutoffs imposed more than six weeks preharvest also caused a small reduction in fruit
number per plant. Declining fruit mass in response to irrigation deficits imposed after
fruit have set has been commonly reported (Cahn et al. , 2003 , 2004; Colla et al. , 1 999;
May and Gonzales, 1 994, 1 999; Mitchell et aI. , 1 991b; Murray, 1 999). Lower fruit
number may have resulted from increased flower abortion under elevated plant water
stress (Atherton and Othman, 1 983; Colla et al. , 1999; PUlupol et al. , 1 996; Wudiri and
Henderson, 1 985) or an increasing number of fruit that rotted and were not counted
(Nichols et al. , 200 1 ; Murray, 1 999). Because most fruit are set between 1 1 -6 weeks
preharvest (fruit set in the final 6 weeks do not typically mature before harvest),
avoiding water stress during this sensitive period appears critical in maximizing fruit
yields.
Although the standard grower practice of irrigating late into the season resulted
in the greatest yield, such practices in drip-irrigated fields have historically caused
highly variable fruit SSC. This variability has been largely related to field-specific
factors (soil texture, rooting depth and wetting pattern away from the drip tape), making
it impossible to develop a generic cutoff date recommendation. An alternative approach
to complete irrigation cutoff is cutback irrigation. Imposing a controlled water deficit
early in the fruit ripening phase has shown the potential to increase SSC with only a
minimal sacrifice in fruit yield (Cahn et al. , 2001 ; Renquist and Reid, 200 1 ). In the
current studies, cutback strategies also appeared to provide this 'middle-ground'
between cutoff dates imposed too early and those imposed too late. In both years, the
yield of moderate cutback deficits (40 d and 50 d) even to 25 % of ETo were
comparable to the standard 20 d irrigation cutoff.
1 34
4.4.4.2 Effect of Water Application on Fruit Soluble Solids
Fruit SSC generally increased with reduced water application during ripening.
This observation was consistent with previous studies (Battilani, 1 994; Cahn et al. ,
200 1 , 2003 ; Calado et al. , 1 990; Colla et al. , 1 999; Lowengart-Aycicegi et aI. , 1 999;
May and Gonzales, 1 994, 1 999). Water stress increases plant water potential (Mitchell
et al. , 1 99 1 a; Renquist and Reid, 200 1 ; Rudich et aI. , 1 98 1 ), which reduces the
movement of solutes into developing fruit; this decreases fruit expansion and results in
higher sugar concentrations (Ehret and Ho, 1 986). While early-season irrigation cutoffs
(60 d or 50 d) increased fruit SSC, they also disproportionately depressed yields and
were therefore uneconomical . Although even the control treatment reached the
commercial benchmark of 5.0 °Brix in both trials, this presumably reflected the modest
yields (high fruit loads can reduce SSC; Dumas et al. , 1 994) and careful management
afforded to experimental plots. In the commercial setting, a large number of drip
irrigated fields have not met this standard using the standard 20 d cutoff approach, most
commonly due to insufficient water stress during fruit ripening. Fruit SSC levels have
been as low as 3.9 °Brix in some drip-irrigated fields (Linden, 2004); although this
represents the extreme, Hartz (2001 ) suggested a consistent decline in fruit SSC of
between 0.2-0.5 °Brix when compared to the furrow average.
To reliably increase fruit SSC, sufficient water stress needs to be imposed
sooner than the standard grower practice achieves; 50-70 % of fruit may have
developed colour by the time irrigation is stopped 20 days preharvest, leaving only a
small percent of green fruit on which subsequent water stresses are effective (Hartz et
al. , 2004). As an alternative to full cutoff, irrigation cutbacks imposed at the onset of
ripening (approximately 6 weeks preharvest) appeared to provide greater flexibility in
managing SSC, as a large percent of fruits were subjected to modest water deficits
without a high risk of excessive yield loss. Across seasons, fruit SSC in treatments that
were cutback to 25 % of ETo at 40 d and 50 d was approximately 0.2 °Brix higher when
compared to the standard 20 d cutoff; this was achieved with only a small yield loss.
Even small improvements in fruit SSC are important to processors because the savings
can be substantial (Renquist and Reid, 2001) ; based on the financial estimations
provided by Linden (2004), an improvement in fruit SSC of 0.2 °Brix would reduce
processing expenses by approximately US$230 per hectare.
1 3 5
4.4.4.3 Effect of Water Application on Brix Yield
Generally, Brix yields did not decline with reduced water application during
fruit ripening. This observation was consistent with the 'dilution' relationship between
fruit yield and SSC, which often results in lower SSC with increasing yield and higher
SSC with decreasing yield (Cahn et al. , 200 1 , 2003 ; Colla et al. , 1 999; Dumas et al. ,
1 994; May and Gonzales, 1 994, 1 999; Phene, 1 999; Renquist and Reid, 200 1 ; Rudich et
al. , 1 977; Stevens and Rudich, 1 978). However, this trade-off relationship was not
unlimited; across years, Brix yields from the earliest cutoff treatments were lower when
compared against only control treatment; average Brix yield loss was approximately 0.5
Mg Brix solids ha-I . This confirmed a disproportionate loss of yield per SSC increase.
At current market pricing in California, the financial value per Mg of Brix solids is
approximately US$ 1 050; decline of 0.5 Mg ha-I could therefore result in a grower loss
of US$525 per hectare. To maximize Brix yields, irrigation cutoff implemented more
than six weeks preharvest should therefore be avoided.
As a practical tool, irrigation cutbacks appeared to provide greater control of
Brix yield than full irrigation cutoff. Irrigation cutback to 25-50 % of ETo imposed at
the onset of fruit ripening resulted in high SSC outcomes with no overall loss in Brix
yield. Although full irrigation cutoff 30 or 40 days preharvest achieved comparable
Brix yields to the standard in the current trials, these strategies do not appear to offer the
same degree of flexibility as irrigation cutback. On lighter-textured soils, irrigation
cutoff at 40 days may have resulted in larger yield loss, while on a heavy-textured soil
irrigation cutoff at 30 days may not have greatly affected fruit SSc. By supplying only
a reduced amount of water during fruit ripening, growers have the beneficial effect of
modest water stress on SSC without the same degree of risk associated with yield loss.
Unlike irrigation cutbacks, cutoff decisions are also typically irreversible, in that by the
time excessive stress conditions are observed, productivity has invariably been lost.
The use of irrigation cutbacks in drip-irrigated fields warranted further research.
In a continuation of these initial trials, irrigation cutback for use in commercial fields
was investigated under conditions representative of the California processing tomato
industry. This research is summarized by Johnstone et al. (2005b, Appendix IV). In addition to irrigation cutback, these trials also advocated monitoring the SSC of early
ripening fruit at a breaker-stage maturity (equivalent to 30-60 % of the fruit surface
showing a colour other than green). Fruit monitoring provides the flexibility to tailor
1 36
irrigation strategies on a field-specific basis; more aggressive water deficit strategies
could be considered in fields with low initial SSC, while modest water deficits could be
used when SSC levels are already high. Because some degree of marketable yield loss
is unavoidable with water stress, predicting the SSC of these early-ripening fruit should
minimize unnecessary loss of productivity.
Cutback techniques may also be promising in humid environments such as N.Z.,
where rainfall can be received during the fruit ripening period. Rainfall can mitigate the
effect of late-season cutoffs (20-30 days), reducing the ability of growers to increase
SSC. A similar problem was described by Burgrnans et al. (1 998); in four of five
seasons, they were unable to impose a substantial water stress on plants because of
rainfall and subsurface water supplies. An irrigation cutback started during early
ripening may therefore allow some degree of moisture stress to accrue with time,
reducing the impact of late rainfall (if received). Such cutback techniques appear to
warrant further research in these types of environments.
4.4.4.4 Effect of Water Application on Blended Fruit Colour
The blended colour of marketable fruit generally improved with reduced water
application during ripening. These results were consistent with those of others (Cahn et
al. , 2003 , 2004; Colla et al. , 1 999; May and Gonzales, 1 999; Zegbe-Dominguez et al. ,
2003); all found that increasing water stress during ripening increased the red colour
intensity of fruit. The relationship between irrigation management and fruit colour is
presumably an indirect result of water stress accelerating fruit ripening (Wolf and
Rudich, 1 988); ripening includes the degradation of green chlorophyll and synthesis of
lycopene.
Importantly, even the colour scores of the control treatments were acceptable for
subsequent processing. Larger treatment differences appeared to be moderated by the
small effect of irrigation on crop maturity. Because all treatments were harvested when
the control treatment was approximately 95 % ripe, most fruit had matured; there was
then an inherently low ratio of green to red colour in all fruit. It is therefore unlikely
that fruit colour will be limited by irrigation management decisions in fields that are
harvested close to commercial maturity.
1 37
4.4.4.5 Effect of Water Application on Fruit pH
Fruit pH was unaffected by reduced irrigation application. This confirmed that
late-season water management does not offer growers a reliable tool to manipulate fruit
pH. Similar observations were made by Cahn et al. (2003); across 1 0 commercial field
experiments, variability was greater between sites than between irrigation treatments.
May and Gonzales ( 1 994) and Murray ( 1 999) also found no consistent relationships
between fruit pH and irrigation. Cultivar selection and harvesting fruit at a correct stage
of maturity remain the most effective strategies for pH manipulation.
4 .4.4.6 Effect of Water Application on Yield Components
There was little effect of irrigation management on fruit maturity; by design,
differences were moderated by selecting a harvest date that reflected the commercial
maturity of the control treatment (i.e. the least stressed treatment). Across seasons, the
percent of ripe fruit at harvest was high (approximately 95 %). Good 'field-storage'
capability in 'Halley' as suggested by May and Guillen (2001 ) meant that the percent of
marketable fruit was similar across most irrigation treatments. Some studies have
suggested that severe irrigation deficits can increase fruit culls and rots (Adegoroye and
Jolliffe, 1 983; Colla et al. , 1 999; May and Gonzales, 1 999; Murray, 1 999). With the
exception of the earliest irrigation cutoffs in each year (60 d and 50 d), all other
treatments resulted in low fruit culls and rots. It is likely that other production
conditions (late rainfall, free moisture on bed tops, humidity and air temperature)
influence fruit culls and rots more severely than irrigation strategies imposed during the
final 6 weeks preharvest. Selecting a harvest date when most fruit are ripe (but before
excessive fruit rots prevail) is therefore important in maximizing marketable yield and
quality. Such a strategy will also minimize the percentage of fruit that are green.
1 38
CHAPTER 5: Summary
The use of drip fertigation has significantly increased the yield of processing
tomatoes, while offering increased fertilizer and irrigation efficiency. Reducing
production inputs may improve overall environmental stewardship. However, current
fertilizer norms may over-estimate plant demand in such systems. Additionally, some
processors have remained hesitant of drip because fruit soluble solids have been lower
than achieved with conventional irrigation strategies. A series of four experiments were
conducted to investigate both fertilizer and irrigation management in drip-irrigated
processing tomatoes. Improved stewardship in both areas not only offers a potential
grower benefit (reduced costs and greater productivity) but also lowers the potential risk
of environmental degradation.
5. 1 1998-1999 Aeroponic Nutrition Trial
A glasshouse nutrition experiment was conducted during the summer of 1 998-
1 999. The experiment was designed to evaluate the effect of low or high fertility
combinations during vegetative development, fruit development and fruit ripening on
plant growth and subsequent fruit yield and quality of processing tomato. Aeroponic
technology was selected to enable sharp comparison of fertility regimes in each growth
period; supply of nutrients directly at the root surface effectively eliminated any
buffering capacity of a soil volume, which has been a limitation of traditional fertilizer
trials in the field. The aeroponic experiment was conducted at Massey University,
Palmerston North. Twelve destructive plant harvests were made to quantify the effect
of time and fertility combinations on plant biomass. Tomato yield and fruit quality
were determined at the final harvest. Cumulative uptake of N, P and K was calculated
for each period of plant growth, and related to maximum yield and fruit quality.
5. 1 . 1 Effect of Fertility on Plant Growth
Seasonal plant biomass accumulation for all treatments was sigmoidal,
consistent with the typical growth habit of processing tomato. Compared to low fertility
treatments, high fertility during early growth increased the vegetative framework
established by the plants. This difference in early vegetative growth largely determined
1 39
subsequent yield outcomes. With high initial fertility, fruit number and total biomass
were maximized, which mostly likely reflected greater assimilate supply in plants that
had established larger vegetative canopies.
Greater accumulation of vegetative biomass with high initial fertility was
attributable to an increased RGR in plants, which itself was related to a higher LAR
early in the season. These observations were consistent with the larger leaf area
developed with high initial fertility; by comparison, low early fertility caused a mild N
deficiency, which reduced leaf area. The P and K status of plants appeared unrelated to
any biomass measure.
5. 1 .2 Effect of Ferti l ity on Leaf Nutrient Concentration s
The N, P and K status of youngest-mature leaves were monitored seasonally and
compared against established nutrient sufficiency norms for processing tomatoes.
5 . 1 .2. 1 Nitrogen Concentration
Seasonally, leaf N concentrations declined as assimilates were redistributed to
developing fruit. In plants receiving low fertility during vegetative development, the
concentration of N in leaf tissue fell increasingly below established sufficiency norms;
tissue values appeared to be mildly deficient. These observations were consistent with
reduced vegetative growth. Early N deficiency appeared to primarily reflect high
demand for this nutrient during the intensive period of vegetative growth. During fruit
development and ripening, tissue N concentrations were above established sufficiency
norms for all treatments.
5 . 1 .2.2 Phosphorus Concentration
Seasonally, the concentration of P in leaf tissue remained very high for all
treatments; values were well above established sufficiency norms. Leaf P concentration
did not decline as fruit developed. Both observations suggested that P availability in
solution was sufficient for optimal growth, even under low fertility; irrespective of
concentration in solution P was continuously available in highly-soluble plant forms.
There was no evidence to suggest that such high concentrations in the plant reflected an
1 40
actual critical growth requirement. Differences in plant growth and fruit yield therefore
appeared unrelated to P status in the plant.
5 . 1 .2.3 Potassium Concentration
Seasonally, the concentration of K in leaf tissue remained very high for all
treatments, well above established sufficiency norms. Given the high early-season K
concentrations in the leaf, there was no evidence to suggest that differences in
vegetative biomass and subsequent fruit yields were related to tissue K status. This
suggested that K availability in solution was sufficient for optimal growth, even under
low fertility. There was no seasonal decline in leaf K concentrations for most
treatments, and in some (particularly those receiving high late fertility), tissue K
concentrations actually increased as fruit developed and ripened. This increase most
likely reflected luxury absorption of K due to availability in solution rather than an
actual plant demand; fruit yields were determined well in advance of such increases.
5. 1 .3 Effect of Fertil ity on Fruit Yield and Qual ity
Maximum total fruit yield was achieved with high initial fertility. Higher yield
was due to greater fruit number rather than mean individual fruit mass, which varied
little between treatments. Marketable yields were low for all treatments because the
final harvest occurred before full maturity. High initial fertility did not increase fruit
culls, nor did it cause a significant delay in crop maturity. Maximizing total yield
would therefore also maximize marketable yield. BER incidence was high across
treatments; there was no evidence to suggest that high fertility regimes increased BER.
Across treatments, fruit SSC was low. Most likely this reflected an earlier than
ideal harvest before full maturity and also a lack of significant osmotic stress during
ripening. There was a small improvement in SSC with high fertility, particularly when
supplied during fruit ripening. It was unclear whether elevated solution conductivity or
higher K availability caused this response. However, the efficacy of high late-season
fertility as fruit ripened appeared poor, because plant nutrient demand after most growth
had ceased was comparatively low. Given the positive effect of high initial fertility
particularly on total yield (and to a lesser extent fruit SSC), Brix yield was also
maximized with high initial fertility. High late-season fertility increased titratable
1 4 1
acidity of fruit, although this improvement was small . There was no effect of fertility
on lycopene content of ripe fruit.
5. 1 .4 Effect of Ferti l ity on Nutrient Uptake
Plant uptake for N, P and K followed closely the pattern of biomass
accumulation. Greatest nutrient uptake was during the period of rapid fruit biomass
accumulation. The fact that nutrient uptake was sigmoidal emphasized that crop
demand for nutrients changes seasonally. Seasonal plant nutrient uptake for high fruit
yield (> 90 Mg ha- I) was equal to 1 2. 1 , 3 .0 and 20.8 g planfl for N, P and K,
respectively; at 22,222 plants ha- I , this was equivalent to approximately 270, 67 and 460
kg ha- I , respectively. Plant N uptake was similar to that in high-yielding commercial
fields, while both P and K were well above typical field levels. Although aeroponic
culture provided the ideal root environment for plant growth and allowed manipulation
of fertility directly at the root surface, the interpretation of nutrient uptake results
obtained from this method appeared confounded by its own unique limitations. All
nutrients were supplied in highly absorbable forms, so although nutrients varied in
concentration, they were readily available. In this system there was luxury uptake of
both P and K.
5.2 1999-2000 Field Nutrition Trial
A field experiment was conducted during the summer growing season of 1 999-
2000, to validate the results found in the aeroponic experiment. Based on the aeroponic
study under controlled conditions, 'optimum' nutrient uptake was quantified for
maximum tomato yield and fruit quality. Closely matching nutrient supply with this
plant demand was tested as a means of maintaining high yield in processing tomato
while improving fertilizer efficiency with small but regular nutrient fertigation. The
experiment was conducted at Massey University, Palmerston North. Soil texture for
this study was a sandy loam. Field treatments compared optimum fertility to rates
below (0 or 0.5x optimum) and above ( 1 .5 and 2.0x optimum) this estimation. Only N
and K treatments were evaluated in the field, because the experimental site had a very
high initial soil test P value. Treatments were supplied by drip fertigation, three times
per week. The nutrient status of the crop was monitored at early bloom, full bloom and
1 42
early fruit ripening. An early destructive plant harvest was made to quantify the effect
of fertility on plant biomass. Tomato yield and fruit quality were determined later in the
season but before commercial maturity.
5.2. 1 Effect of Ferti l ity on Leaf Nutrient Concentrations
The N, P and K status of youngest-mature leaves were monitored seasonally and
compared against established nutrient sufficiency norms for processing tomatoes.
5 .2 . 1 . 1 Nitrogen Concentration
As in the aeroponic experiment, leaf N concentration declined seasonally as
assimilates were redistributed to fruit. Optimal fertility and above resulted in tissue N
values interpreted as sufficient for high yield. In plants receiving above optimal fertility
tissue N concentrations were not supra-optimal despite very large N applications (as
much as double typical commercial applications); this suggested plant demand was not
as high as these rates supplied. Tissue N concentration appeared to be mildly deficient
during early bloom in treatments receiving less than optimal fertility; a similar
observation was made for the low fertility treatment under aeroponics. Late-season
tissue N concentration in these treatments was not below established sufficiency values,
suggesting that N application after most active fruit growth has stopped was
unnecessary.
5.2. 1 .2 Phosphorus Concentration
All treatments resulted in adequate tissue P status compared to established
sufficiency norms. This reflected the high soil test P value at the experimental site.
Leaf P concentration declined seasonally in the field. Tissue P concentrations were
considerably lower than found under aeroponics, even at what was considered to be a
high soil test P level. This confirmed that there was luxury uptake of P in the
aeroponics experiment.
1 43
5.2 . 1 .3 Potassium Concentration
All treatments resulted in adequate to high tissue K status compared to
established sufficiency norms. Seasonal decline was observed in all treatments, even
those supplying large amounts of fertilizer K; this reflected redistribution of K to the
fruit. Although K application during fruit ripening resulted in greater leaf K
concentrations, such applications were unnecessary for maximum yield and fruit
quality. Field results confirmed that the high late-season tissue K concentrations under
aeroponics represented luxury levels; K uptake under glasshouse conditions therefore
over-estimated actual plant demand.
5.2.2 Effect of Ferti l ity on Plant Growth
Plants receiving optimal fertility and above established the greatest total biomass
and leaf area by flowering. The N status of plants was most important in accounting for
these differences; plants with low biomass and small leaf area had leaf tissue N
concentration below sufficiency norms at early bloom. The P and K status of plants
were adequate for all treatments. These findings were consistent to those made under
aeropomcs.
5.2 .3 Effect of Fertility on Fruit Yield and Quality
High total fruit yield was achieved with treatments that supplied optimal fertility
and above. Higher yield was due primarily to greater fruit number rather than mean
fruit mass. Improved fruit number appeared associated with the larger vegetative
frameworks established early in the season. These findings were consistent to those
made under aeroponics, and were also related to early-season N status in plants rather
than to P or K. Marketable yields were not maximized for any treatment; this reflected
an early harvest date due to an increasing incidence of plant disease and fruit rots.
Above-optimal fertility resulted in the lowest percent of total yield that was marketable.
This was linked to high tissue N status which delayed maturity. However, there was no
effect of fertility on fruit culls. If all treatments had been grown to commercial
maturity, marketable yields would have been greatest in plants receiving optimal
144
fertility and above. The incidence of BER was almost non-existent In the field
compared to levels found under glasshouse aeroponic culture.
Applying above optimal fertilizer did not increase fruit yields compared to the
optimal treatment. The likelihood of N (and possibly K) leaching out of the active root
zone was high due to heavy summer rainfall and a light-textured soil. It appeared that
the optimal fertility treatment itself would most likely have been sufficient to maximize
growth and subsequent yields under normal field conditions (i.e. a drier season).
Fertilizer applications made during fruit ripening were not useful, as mean fruit mass
was not improved. Maximum yields therefore appeared attainable at rates equivalent to
plant uptake determined for only the initial two periods of growth under aeroponics
(vegetative development and fruit development); at a density of 22,222 plants ha- 1 this
was equal to approximately 2 1 0 and 3 1 0 kg ha- 1 of N and K, respectively. These rates
were comparable with current supply in commercial processing tomato fields.
Across treatments, low fruit SSC reflected high soil moisture availability during
ripening and also a final harvest before full maturity. Higher fertility resulted in greater
fruit SSC. This improvement may have been associated with the larger plant canopies
established with higher fertilizer rates; larger canopies often have improved assimilate
capacity. There was no correlation between fruit tissue K concentration and fruit SSc.
Application of K fertilizer during fruit ripening was therefore neither a reliable nor cost
effective means to favourably influence SSC, and did not appear to offer any further
benefit beyond appropriate late-season water management; considerably higher fruit
SSC levels were achieved without excessive K application but where late-season soil
moisture stress was imposed by rain-shelter. Given the positive effect of at least
optimal fertility on total yield and fruit SSC, Brix yield was also maximized for these
treatments.
The concentration of N, P and K in fruit reflected the overall availability of each
nutrient in the plant, but otherwise appeared unimportant. Nutrient removal in the
harvested marketable product was closely associated with fruit yields; higher yields
resulted in greater removal of N, P and K. However, the majority of applied N and K
were not removed in the harvested fruit, and therefore either remained in the soil, was
leached or returned to the soil as crop residue. Fertilizer efficiency was poor, less than
25 % of applied N and K for all fertilizer treatments. The very low efficiency at above
optimal fertility confirmed that such applications were not only unjustified, but
extremely wasteful; these applications exceeded plant demand under normal field
1 45
conditions. Despite poor efficiency, fertigation technology still offered the potential to
closely match plant demand with supply compared to conventional pre-plant and side
dressing fertilization techniques.
5.3 2001 and 2002 UCD Irrigation Management Trials
Management for high fruit SSC levels was not reliably achieved by heavy late
season fertilization. Subsequently, irrigation experiments were conducted during the
summer of 2001 and again in 2002. The aim was to evaluate the effects of irrigation
management during fruit ripening on yield and SSC of drip-irrigated processing
tomatoes. Late-season irrigation cutoff and cutback treatments were investigated, from
no water during the final 60 days preharvest (late fruit growth) to no water during the
final 20 days preharvest (the current commercial standard). These experiments were
conducted at the University of California, Davis. California has an arid environment,
ideal for irrigation trials . Soil texture in both years was a loam. Plant canopy cover,
plant water potential, and soil water status were documented during fruit ripening in
response to the irrigation deficits imposed. At commercial maturity, fruit yield and
quality were assessed.
5.3. 1 Effect of Water Application on Plant Canopy Cover
Plant canopy cover decreased seasonally with natural plant senescence. A high
tolerance to water stress without canopy decline was observed; even the most severe 60
d irrigation cutoff did not accelerate plant die-back or vine collapse. This degree of
restriction to the water supply may have caused canopy dieback in fields with a lighter
textured soil type and low moisture retention. Cultivar selection may also affect canopy
dieback.
5.3.2 Effect of Water Appl ication on Plant Water Status
5 .3 .2 . 1 Effect of Sampling Technique
Bagging leaves overnight for early morning sampling the next day extended the
period during which estimates of predawn plant water potential could be made. This
1 46
would allow growers to monitor a greater number of fields between 0500 HR (sunrise)
and 0700 HR; previously all measurements had to be completed before sunrise, which
limited commercial application of the technique. Bagging leaves for at least two hours
before afternoon measurement reduced sampling variability related to transient
environmental factors. By substantially reducing transpiration, the bagging technique
allowed the water potential of leaves to equilibrate with the stem and roots. This
provided a more reliable indicator of overall plant water status as related to fruit yield
and quality.
5 .3 .2.2 Effect of Irrigation Treatment
Early morning and early afternoon measurements of plant water potential on
bagged leaves confirmed that irrigation treatments caused varying levels of water stress.
Generally, the values measured reflected the percent of ETo applied; greater irrigation
amounts resulted in lower plant water stress. Average plant water potential during fruit
ripening was consistently correlated to fruit yield and SSC at maturity; measurements
made in the early afternoon typically resulted in the strongest relationships with yield
and quality. More negative plant water potential (indicative of greater water stress) was
associated with lower yield and higher fruit SSC. Average 1 300 HR '!'Stem in the range of
-0.5 to -0.7 MPa during the final six weeks preharvest was sufficient to achieve fruit
SSC levels > 5.0 °Brix.
5.3.3 Effect of Water Appl ication on Soil Water Status
Measurements of Bv confirmed that irrigation practices achieved varying levels
of water availability within the soil. Generally, increasing the percent of ETo applied
increased soil water availability. Average By during fruit ripening was not as strongly
correlated to yield and fruit SSC as 1 300 HR '!'Stem. No reliable '!'Soil target during fruit
ripening was found to achieve acceptable SSC.
1 47
5. 3.4 Effect of Water Appl ication on Yield and Qual ity
5 .3 .4. 1 Total and Marketable Yield
Total and marketable yield generally declined with decreasing water application
during fruit ripening. Lower yield was primarily related to smaller fruit size, although
severe irrigation cutoffs imposed more than six weeks preharvest also caused a small
reduction in fruit number per plant. Lower fruit number may have resulted from
increased flower abortion under elevated plant water stress or an increasing number of
fruit that rotted and were not counted. Irrigation strategies should prevent water stresses
that impede fruit set or that cause a substantial increase in rots; this precluded early
season cutoffs implemented more than six weeks preharvest. Although the standard
grower practice of irrigating late into the season resulted in the greatest yield, this
strategy did not maximize fruit Ssc. Irrigation cutback strategies implemented during
early ripening provided an intermediate degree of water stress, and balanced yield
decline with a proportional SSC increase.
5 .3 .4.2 Fruit Soluble Solids
Fruit SSC generally increased with reduced water appiication during ripening.
To reliably increase fruit SSC, sufficient water stress needs to be imposed sooner than
the standard grower practice achieves; 50-70 % of fruit may have developed colour by
the time irrigation is stopped 20 days preharvest, leaving only a small percent of green
fruit on which subsequent water stresses are effective. While early-season cutoffs
increased SSC they disproportionately depressed yields. As an alternative to cutoffs,
irrigation cutbacks imposed at the onset of ripening provide considerable flexibility in
managing SSC, as a large percent of fruits can be subjected to modest water deficits.
5 .3 .4.3 Brix Yield
Generally, Brix yields did not decline with reduced water application during
fruit ripening. This was consistent with the inverse relationship observed between fruit
yield and SSC. However, this trade-off relationship was not unlimited; applying less
than 20 % of ETo during the final 60 days resulted in a significantly lower Brix yield.
1 48
To maximize Brix yields, irrigation cutoff implemented more than six weeks preharvest
should be avoided. As a practical tool, irrigation cutbacks provide greater control of
Brix yield than full irrigation cutoff; an irrigation cutback to 25-50 % of ETo imposed at
the onset of fruit ripening resulted in high SSC outcomes with no overall loss in Brix
yield. In a series of later commercial trials, monitoring the SSC of early ripening fruit
at a breaker-stage maturity (equivalent to 30-60 % of the fruit surface showing a colour
other than green) was a useful tool to guide irrigation cutback. Fruit monitoring
provided the flexibility to tailor irrigation strategies on a field-specific basis; more
aggressive water deficit strategies could be considered in fields with poor initial SSC,
while modest water deficits could be used when SSC levels are already high.
5 .3 .4.4 Blended Fruit Colour
The blended colour of marketable fruit generally improved with reduced water
application during ripening. However, the colour scores of even the control treatments
were acceptable for subsequent processing. It is unlikely that fruit colour will be
limited by irrigation management decisions in fields that are harvested close to
commercial maturity.
5 .3 .4.5 Fruit pH
Fruit pH was unaffected by reduced irrigation application. This confirmed that
late-season water management does not offer growers a reliable tool to manipulate fruit
pH. Cultivar selection and harvesting fruit at a correct stage of maturity remain the
most effective strategies for pH manipulation.
5 .3 .4.6 Yield Components
There was little effect of irrigation management on fruit maturity; by design,
differences were moderated by selecting a harvest date that reflected the commercial
maturity of the standard grower cutoff. Good 'field-storage' capability meant that the
percent of marketable fruit was similar across most irrigation treatments; culled or
rotten fruit accounted for only a small percent of total yield. Selecting a harvest date
1 49
when most fruit are ripe but before excessive fruit rots prevail IS important in
maximizing yield and quality_
1 50
CHAPTER 6: Conclusions
6. 1 Plant Nutrition Studies
.:. Nutrition studies confinned that growers should maximize vegetative growth
during the 6-8 weeks after transplanting. Nutrient supply should not limit
vegetative growth during this period of development; adequate N supply is
particularly important in establishing a large and healthy canopy. Maximizing
vegetative growth was necessary to maximize fruit set. Reinstating adequate
fertility after vegetative growth had stopped and fruit number had been
detennined did not improve yield. High fruit quality was not strongly related to
K nutrition. In particular, large late-season K applications to improve fruit SSC
did not offer any further benefit beyond appropriate late-season water
management. Collectively, these findings were largely consistent with existing
literature .
• :. Aeroponic and field studies confinned that plant nutrient demand changes
seasonally. Although adequate fertility is required for vegetative growth,
nutrient uptake was greatest during fruit development; where practical, fertilizer
application should be concentrated during this period of growth. Applying
fertilizers during fruit ripening was not justified for maximum fruit yield and
quality. Closely matching nutrient supply with plant demand may improve
fertilizer efficiency compared to traditional pre-plant and side-dressed
applications. Limiting fertilizer application during periods when demand is low
will minimize nutrient loss potential. Drip fertigation offers an effective means
in which to match supply more closely with crop uptake. Seasonal plant uptake
equal to 9.4 and 1 3 .8 g planfl of N and K, respectively, was sufficient to
maximize vegetative growth and achieve high fruit yield (> 90 Mg ha- I ) ; at a
medium density of 22,222 plants ha-I this was equivalent to approximately 2 1 0
and 3 1 0 kg ha-I of N and K, respectively. These numbers were consistent with
current commercial N and K applications; higher rates would only be justified
where field-specific factors dictated.
1 5 1
.:. Closely matching supply with periods of most intensive nutrient uptake by
plants and accurately predicting the availability of nutrients already in the soil
offer good potential for growers to reduce fertilizer applications and improve
environmental stewardship. Fertility status can vary greatly from site-to-site,
and so management tools that allow fertilizer decisions to be based on robust
estimates of nutrient availability in the soil (as related to fruit yield and quality)
appear particularly promising.
6. 1 Irrigation Management Studies
.:. Irrigation cutoff prior to fruit ripening can reduce fruit set, decrease fruit size,
and substantially increase the incidence of fruit rots; growers should therefore
minimize plant water stress prior to the final six weeks preharvest. However,
late-season irrigation deficits applied during fruit ripening can increase fruit SSC
without excessive yield loss. Irrigation cutbacks provide greater grower control
of Brix yield outcomes than full irrigation cutoff; an irrigation cutback to 25-50
% of ET 0 imposed at the onset of fruit ripening appears sufficient to improve
fruit SSC and maintain Brix yields as compared to the current industry practice
(late irrigation cutoff) .
• :. Brix monitoring of the earliest ripening fruit (when 30-60 % of the fruit surface
shows a colour other than green) can help classify fields as to the severity of
irrigation cutback required to reach desirable fruit SSC at harvest. More
aggressive water deficit strategies could be considered in fields with poor initial
SSC, while modest water deficits could be used when SSC levels are already
high. Other important quality determinants such as fruit colour and pH are
unlikely to be adversely affected by such irrigation practices. Similarly,
irrigation cutbacks initiated during fruit ripening are unlikely to cause extreme
canopy dieback, so there is not a greatly increased risk of increased fruit culls
due to excessive exposure of fruit to direct sun and heat.
.:. Compared to traditional exposed-leaf measures, bagging healthy leaves to
determine plant water status extended potential sampling windows and reduced
sampling variability. Even with this technique, measuring plant water status
1 52
remams an unappealing option to many commercial growers because of the
considerable effort and user-expertise required. Soil-based measures of water
content continue to be a useful management tool to avoid over- or under
watering, and the equipment can be automated or permanently set in fields.
1 53
LITERATURE CITED:
Adams, P. 1 986. Mineral nutrition, p.28 1 -334. In: J.G. Atherton and J. Rudich (eds.).
The tomato crop: a scientific basis for improvement. Chapman and Hall, New
York, N.Y.
Adams, P . , J.N. Davies and G.W. Wins or. 1 978. Effects of nitrogen, potassium and
magnesium on the quality and chemical composition of tomatoes grown in peat.
J. Hort. Sci. 53: 1 1 5 - 122 .
Adams, P and L.e. Ho. 1 992. The susceptibility of modem tomato cultivars to blossom
end rot in relation to salinity. J. Hort. Sci. 67:827-839.
Adams, P. and L.C. Ho. 1993. Effects of environment on the uptake and distribution of
calcium in tomato and on the incidence of blossom-end rot. Plant Soil 1 54 : 1 27-
1 32 .
Adegoroye, AS. and P.A. Jolliffe. 1 983. Initiation and control of sunscald injury of
tomato fruit. J . Amer. Soc. Hort. Sci . , 1 08 :23-28.
Amable, R.A and S. Sinnadurai. 1 977. The influence of potassium, calcium and
irrigation treatments on tomato fruit quality. Acta Hort. 53 : 1 65-1 70.
Andrews, W.F. 1 972. Soil survey of Yolo County, California. D.S. Dept. Agr. Soil
Conserv. Serv. , Washington, D.C.
Atherton, J.G. and G.P. Harris . 1 986. Flowering, p. 1 67-200. In: J .G. Atherton and J.
Rudich (eds.). The tomato crop: a scientific basis for improvement. Chapman
and Hall, New York, N.Y.
Atherton, J.G. and S. Othman. 1 983. Effects of irradiance and water stress on flower
abortion in the glasshouse tomato. Acta Hort. 1 34: 1 33- 1 38 .
Baldet, P. , e. Devaux, C. Chevalier, R. Brouquisse, D. Just and P. Raymond. 2002 .
Contrasted responses to carbohydrate limitation in tomato fruit at two stages of
development. Plant Cell Environ. 25 : 1 639- 1 649.
Barak, J., J.D. Smith, AR. Krueger and L.A. Peterson. 1 996. Measurement of short
term nutrient uptake rates in cranberry by aeroponics. Plant Cell Environ.
1 9:237-242.
Barrett, D.M. and G. Anthon. 200 l . Lycopene content of California-grown tomato
varieties. Acta Hort. 542 : 1 65- 1 73 .
Bartholomew, R.P. and G . Janssen. 1 929. Luxury consumption of potassium by plants
and its significance. J. Amer. Soc. Agron. 2 1 :75 1 -765 .
1 54
Battilani, A. 1 994. Effects of irrigation, ethrel and harvest date on yield and quality of
processing tomato. Acta Hort. 376:299-302.
Besford, R.T. and G.A. Maw. 1 975. Effect of potassium nutrition on tomato plant
growth and fruit development. Plant and Soil 42 :395-4 1 2.
Biddinger, E.J., C. Liu, R.J. Joly and KG. Raghothama. 1 998. Physiological and
molecular responses of aeroponically grown tomato plants to phosphorus
deficiency. J. Amer. Soc. Hort. Sci. 1 23 :330-333.
Blackmer, A.M. 1 987. Losses and transport of nitrogen from soils, p . 85- 1 03 . In: F.M.
D'Itri and L.G. Wolfson (eds.). Rural groundwater contamination. Lewis Publ.
Inc. Chelsea, MI.
Blakemore, L.C., P.L. Searle and B.K Daly. 1 987. Methods for chemical analysis of
soils. N.Z. Soil Bur. Scientific Rpt. 80. N.Z. DSIR, Wellington, N.Z.
Brady, N.C. 1 990. The nature and properties of soils. 1 0th ed. Macmillan, New York,
N.Y.
Bruggink, G.T. and E. Heuvelink. 1 987. Influence of light on the growth of young
tomato, cucumber and sweet pepper plants in the greenhouse: effects on relative
growth rate, net assimilation rate and leaf area ratio. Scientia Hort. 3 1 : 1 61 - 1 74.
Burgess, S.M. 1 988. The climate and weather of Manawatu. N.Z. Meteorological
Service Pub. 1 1 5( 1 8) . Wellington, N.Z.
Burgmans, J .L., J.P. C. Watt and W.T. Bussel1. 1 998. Lcssons from processing tomato
irrigation trials in Hawkes Bay, New Zealand. N.Z. 1. Crop Hort. Sci. 26:307-
3 1 2.
Cahn, M., B. Hanson and R. Snyder. 2000. Weather stations for irrigation management,
p. 1 1 - 1 8 . California Tomato Research Institute Annual Project Reports. CTRI
Inc. , Escalon, CA.
Cahn, M.D., E.V. Herrero, B.R. Hanson, T.K Hartz and E.M. Miyao. 2001 . Water
management strategies for improving fruit quality of drip-irrigated processing
tomatoes. Acta Hort. 542: 1 1 1 - 1 1 6.
Cahn, M.D., E.V. Herrero, B.R. Hanson, T.K Hartz and E.M. Miyao. 2004. Late
season water management effects on quality and yield of processing tomatoes.
Acta Hort. 664: 1 1 1 - 1 1 8 .
Cahn, M.D., E.V. Herrero, B.R. Hanson, R.L. Snyder, T.K Hartz and E.M. Miyao.
2003 . Effects of irrigation cutoff on processing tomato fruit quality. Acta Hort.
6 1 3 :75-80.
1 55
Calado, AM., A Monzon, D.AClark, C.J. Phene, C. Ma and Y. Wang. 1 990.
Monitoring and control of plant water stress in processing tomato. Acta Hort.
277 : 1 29- 1 36.
Carangal, AR., E.K. Alban, J.E. Vamer and RC. BUffell. 1 954. The influence of
mineral nutrition on the organic acids of the tomato, Lycospersicon esculentum.
Plant Phys. 29:355-360.
Cassman, K.G., D.C. Bryant, and B.A Roberts. 1 990. Comparison of soil test methods
for predicting cotton response to soil and fertilizer potassium on potassium
fixing soils. Comm. Soil Sci. Plant Anal. 2 1 : 1 727- 1 743.
Cavero, J. , R.E. Plant, C. Shennan, and D.B. Friedman. 1 997. The effect of nitrogen
source and crop rotation on the growth and yield of processing tomatoes. Nutr.
Cycl. Agroecosyst. 47:271 -282.
Cavero, J., RE. Plant, C. Shennan, and D.B. Friedman, J.R. Williams, lR Kiniry and
V. W. Benson. 1 999. Modelling nitrogen cycling in tomato-saftflower and
tomato-wheat rotations. Agr. Syst. 60: 123- 1 3 5 .
Cavero, J. , R.E. Plant, C . Shennan, J.R. Williams, J.R Kiniry and V.W. Benson. 1 998.
Application of EPIC model to nitrogen cycling in irrigated processing tomatoes
under different management systems. Agr. Syst. 56:391 -4 1 4.
Cerda, A., F.T. Bingham, G.J. Hoffman and c.K. Kuszar. 1 979. Leaf water potential
and gaseous exchange of wheat and tomato as affected by NaCi and P leveis in
the root medium. Agron. J. 7 1 :27-3 1 .
Ceme, M . 1 990. Different kinds and levels of nitrogen nutrition in tomatoes. Acta Hort.
277 : 1 79-1 82.
Clark, M.S. , W.R Horwath, C. Shennan, K.M. Scow, W.T. Lanini and H. Ferris. 1 999.
Nitrogen, weeds and water as yield-limiting factors in conventional, low-input,
and organic tomato systems. Agric. Ecol. Environ. 73 :257-270.
Colla, G., R Casa, B. Lo Cascio, F. Saccardo, O. Temperini and C. Leoni. 1 999.
Responses of processing tomato to water regime and fertilization in central Italy.
Acta Hort. 487:53 1 -535.
Colla, G., A Battistelli, S. Moscatello, S . Proietti, R. Casa, B. Lo Cascio and C. Leoni.
200 1 . Effects of reduced irrigation and nitrogen fertigation rate on yield,
carbohydrate accumulation, and quality of processing tomato. Acta Hort.
542 : 1 87-1 96.
1 56
Cuartero, J. and R. Femandez-Munoz. 1999. Tomato and salinity. Scientia Hort. 78:83-
1 25 .
Dadomo, M. , AM. Gainza, Y. Dumas, P. Bussieres, J.I. Macua, M. Christou and X.
Branthome. 1 994a. Influence of water and nitrogen availability on yield
components of processing tomato in the European union countries. Acta Hort.
376:271 -274.
Dadomo, M., S. Leoni, A Rodriguez, T. Koutsos, J.I. Macua, M.L. Taborda, C.M.
Portas and G. Gandolfi. 1 994b. 1 992 Main results of the EEC research
programme - action C. variability in relation to cultivar and country. Acta Hort.
376:43-50.
Davies, J.N. 1 964. Effect of nitrogen, phosphorus and potassium fertilizers on the non
volatile organic acids of tomato fruit. J. Sci. Food Agric. 1 5 :665-673 .
Davies, J.N. and G.E. Hobson. 1 98 1 . The constituents of tomato fruit - the influence of
environment, nutrition and genotype. CRC Critical Rev. Food Sci. Nutr. 1 5 :205-
280.
Davies, J.N. and G.W. Winsor. 1 967. Effect of nitrogen, phosphorus, potassium,
magnesium and liming on the composition of tomato fruit. J. Sci. Food Agric.
1 8 :459-466.
Davies, J.N. and G.W. Winsor. 1 969. Some effects of variety on the composition and
quality of tomato fruit. J. Hort. Sci. 44:33 1 -342.
Dumas, Y. 1 990. Tomatoes for processing in 90's : nutrition and crop fertilization. Acta
Hort. 277 : 1 55- 166.
Dumas, Y., C. Leoni, C.AM. Portas and B. Bieche. 1 994. Influence of water and
nitrogen availability on yield and quality of processing tomato in the European
Union countries. Acta Hort. 376: 1 85- 1 92.
Ehret, D.L. and L. Ho. 1 986. The effects of salinity on dry matter partitioning and fruit
growth in tomatoes grown in nutrient film culture. J. Hort. Sci. 6 1 :361 -367.
Eymar, E., C. Cadahia and A Sanchez. 200 1 . Foliar nutrient reference levels obtained
in hydroponic cultures as preliminary norms for DRIS to fertigate conifers.
Commun. Soil Sci. Plant Anal. 32:267-282.
Fereres, E., D.A Goldhamer and L.R. Parsons. 2003. Irrigation water management for
horticultural crops. HortScience 38 : 1 036- 1 042.
Fisher, K.J . , P.R. Johnstone and M.A. Nichols. 2002. Nutrition of processing tomato.
Acta Hort. 571 :45-49.
1 57
Flores, P. , J.M. Navarro, C. Garrido, J.S. Rubio and V. Martinez. 2004. Influence of
Ca2+, K+ and N03- fertilization on nutritional quality of pepper. J. Sci. Food
Agric. 84:569-574.
Fontes, P.C.R and G.E. Wi1cox. 1 984. Growth and phosphorus uptake by tomato
cultivars as influence by phosphorus concentrations in soil and nutrient solution.
J. Amer. Soc. Hort. Sci. 1 09:633-636.
Fulton, A, R Buchner, B. Olson, L. Schwankl, C. Gilles, N. Bertagna, J. Walton and K.
Shackel. 200 1 . Rapid equilibration of leaf and stem water potential under field
conditions in almonds, walnuts, and prunes. HortTechnology 1 1 :609-61 5.
Geisenberg, C. and K. Stewart. 1 986. Field crop management, p . 5 1 1 -557. In: J.G.
Atherton and J . Rudich (eds.). The tomato crop: a scientific basis for
improvement. Chapman and Hall, New York, N.Y.
Gillaspy, G., H. Ben-David and W. Gruissem. 1 993 . Fruit: a developmental perspective.
Plant Cell 5 : 1439- 145 1 .
Goldhamer, D.A and RL. Snyder. 1 989. Irrigation scheduling: a guide for efficient on
farm water management. Univ. of Calif. Coop. Ext. Bul. 2 1 454.
Greenwood, D.J., F. Gastal, G. Lemaire, A Draycott, P. Millard and J.J. Neeteson.
1 99 1 . Growth rates and %N of field grown crops: theory and experiments. Ann. Bot. 67: 1 8 1 - 190.
Greenwood, D.J. 200 1 . Modeling N-response of field vegetable crops grown under
diverse conditions with N ABLE: a review. J. Plant Nutr. 24: 1 799- 1 8 1 5 .
Grierson, D. and AA. Kader. 1986. Fruit ripening and quality, p.241 -280. In: J.G.
Atherton and J. Rudich (eds.). The tomato crop: a scientific basis for
improvement. Chapman and Hall, New York, N.Y.
Giiler, S . , H. Ibrikci, A. Apaydin and H. Kar. 2002. Response of field-grown tomatoes
to nitrogen and potassium applied with drip or furrow irrigation. Acta Hort.
571 : 1 87- 1 93 .
Halbrooks, M.C. and G.E. Wi1cox. 1 980. Tomato plant development and elemental
accumulation. J. Amer. Soc. Hort. Sci. 1 05 :826-828.
Hanson, B.R 1 995. Practical potential irrigation efficiencies. Proc. 1 st Int. Conf. Water
Resources Engineering Div., Amer. Soc. Civil Engineering, San Antonio, Tex.
1 4- 1 8 Aug., 1 995.
Hanson, B.R., D.M. May and L.J. Schwankl. 2003 . Effect of irrigation frequency on
subsurface drip irrigated vegetables. HortTech. 1 3 : 1 1 5- 1 20.
1 58
Hanson, B.R., S . OrIoff and D. Peters. 2000a. Monitoring soil moisture helps refine
irrigation management. Calif. Agr. 54(3) :38-42.
Hanson, B.R. and D. Peters. 2000. Soil type affects accuracy of dielectric moisture
sensors. Calif. Agr. 54(3) :43-47.
Hanson, B.R., D. Peters and S . Odoff. 2000b. Effectiveness of tensiometers and
electrical resistance sensors varies with soil conditions. Calif. Agr. 54(3) :47-50.
Hartz, T. 1 996. Water management in drip-irrigated vegetable production.
HortTechnology 6: 1 65-1 67.
Hartz, T.K. 200 1 . Status of drip irrigation for processing tomato production in
California. Acta Hort. 542 : 1 0 1 - 1 02.
Hartz, T.K. 2002a. Sustainable vegetable production in California: current status, future
prospects. HortScience 37 : 1 0 1 5- 1 022.
Hartz, T.K. 2002b. Eady season irrigation of processing tomato. Vege. Notes, Tulare
and Kings Counties, April 2002. Univ. Calif. Cooperative Ext., Tulare, CA.
Hartz, T.K. 2005. Vegetable production best management practices to minimize nutrient
loss. HortScience (In press.).
Hartz, T.K. and G.J. Hochmuth. 1 996. Fertility management of drip-irrigated vegetable.
HortTechnology 6(3) : 1 68-1 72.
Hartz, T.K. and P.R. Johnstone. 2005 . Relationship between soil P availability and P
loss potentiai in runoff and ieaching. Camm. Soil Sci. Plant Anal. (11 press.).
Hartz, T.K., P.R. Johnstone, D.M. Francis and E.M. Miyao. 2005 . Processing tomato
yield and fruit quality improved with potassium fertigation. HortScience
40: 1 862- 1 867.
Hartz, T.K., P.R. Johnstone and M. LeStrange. 2004. Late-season deficit irrigation is a
reliable tool for soluble solids improvement in drip-irrigated processing tomato
production. Acta Hort. (In press.).
Hartz, T.K. and G. Miyao, 1 997. Processing tomato production in California. Division
of Ag. and Nat. Res. Publ. 7228. Univ. Calif. , Oakland, CA.
Hartz, T.K., E.M. Miyao, R.J. Mullen and M.D. Cahn. 200 1 . Potassium fertilization
effects on processing tomato yield and fruit quality. Acta Hort. 542: 1 27- 1 34.
Hartz, T.K., G. Miyao, R.J. Mullen, M.D. Cahn, J. Valenica and K.L. Brittan. 1 999.
Potassium requirements for maximum yield and fruit quality of processing
tomato. J. Amer. Soc. Hort. Sci. 1 24 : 1 99-204.
1 59
Hartz, T.K., E.M. Miyao and J.G. Valencia. 1 998. DRIS evaluation of the nutritional
status of processing tomato. HortScience 33 :830-832
Hills, F.J., F.E. Broadbent and O.A. Lorenz. 1 983. Fertilizer nitrogen utilization by
corn, tomato and sugarbeet. Agron. J. 75 :423-426.
Hinds, A.A. and L.E. Lowe. 1 980. Application of the Berthelot reaction to the
determination of ammonium-N in soil extracts and soil digests. Comm. Soil Sci.
Plant Anal. 1 1 :469-475.
Ho, L.e. 1 984. Partitioning of assimilates In fruiting tomato plants. Plant Growth
Regulation 2 :277-285 (Abstr.).
Ho, L.e. R I. Grange and AJ. Picken. 1 987. An analysis of the accumulation of water
and dry matter in tomato fruit. Plant Cell Environ. 10 : 1 57 - 1 62.
Ho, L.C. and J .D. Hewitt, 1 986. Fruit development, p.201 -239. In: J.G. Atherton and J .
Rudich (eds.). The tomato crop: a scientific basis for improvement. Chapman
and Hall, New York, N.Y.
Hochmuth, G., D. Maynard, C. Vavrina and E. Hanlon. 1 99 1 . Plant tissue analysis and
interpretation for vegetable crops in Florida. Florida Coop. Ext. Spec. Ser. SS
VEC-42.
Hochmuth, G.J., O. Carrijo and K. Shuler. 1 999. Tomato yield and fruit size did not
respond to P fertilization of a sandy soil testing very high in Melich- l P .
HortScience 34:653-656.
Johnstone, P.R, T.K. Hartz, M.D. Cahn and M.R Johnstone. 2005a. Lettuce response
to phosphorus fertilization in high P soils. HortScience 40: 1499- 1 503 .
Johnstone, P.R, T.K. Hartz, M. LeStrange, JJ. Nunez and E.M. Miyao. 2005b.
Managing fruit soluble solids with late-season deficit irrigation in drip-irrigated
processing tomato production. HortScience 40: 1 857 - 1 86 1 .
Julian, A.P. 1 990. Flower and fruit development in processing tomato. A thesis
presented in partial fulfillment of the requirements for the degree of Master of
Horticultural Science in Vegetable Production at Massey University.
Palmerston North, N.Z.
Keeney, D.R and RF. Follett. 1 99 1 . Managing nitrogen for groundwater quality and
farm profitability: overview and introduction, p. I -7. In: RF. Follett et al. (eds.).
Managing nitrogen for groundwater quality and farm profitability. SSSA, Inc.
Madison, WI.
1 60
Klute, A 1 986. Water retention: laboratory methods, p.635-662. In: A Klute (ed.).
Methods of soil analysis: part 1 . physical and mineralogical methods. ASA,
Madison, WI.
Krusekopf, H.H., J.P. Mitchell, T.K. Hartz, D.M. May, E.M. Miyao and M.D. Cahn.
2002. Pre-sidedress soil nitrate testing identifies processing tomato fields not
requiring sidedress N fertilizer. HortScience 37:520-524.
Lachover, D. 1 972. The effect of potassium on "Rom a" variety of processing tomato,
with special reference to potassium uptake, yield and quality. Qual. Plant Mater.
Veg. XXI 3 : 1 65- 1 77.
Leathwick, J.R., G. Wilson and R.T.T. Stevens. 2002. Climate surfaces for New
Zealand. Landcare Research Contract Report #LC9798/1 26. Manaaki Whenua -
Landcare Research, Hamilton, New Zealand.
Linden, T. 2003. Drip irrigation: getting another look at processing tomato fields. Calif.
Tomato Grower 46(3) :8- 1 3 .
Linden, T. 2004. The brix discussion. Calif. Tomato Grower 47(2) :5-8.
Lingle, J.C, and O.A. Lorenz. 1 969. Potassium nutrition of tomatoes. J. Amer. Soc.
Hort. Sci. 94:679-(j83.
LOpez, J. , R.M. Ruiz, R. Ballesteros, A Ciruelos and R. Ortiz. 200 1 . Color and
lycopene content of several commercial tomato varieties at different harvesting
dates. Acta Hort. 542:243-247.
Lorenz, O.A and K.B. Tyler. 1 983. Plant tissue analysis of vegetable crops (pp. 24-29) .
In H.M. Reisenauer (ed.) Soil and plant tissue testing. California Coop. Ext.
Bul. 1 879.
Lowengart-Aycicegi, A. , H. Manor, R. Krieger and G. Gera. 1 999. Effects of irrigation
scheduling on drip-irrigated processing tomatoes. Acta Hort. 487: 5 1 3-5 1 8.
Maas, E. V. 1 986. Salt tolerance of plants. Appl. Agric. Res. 1 : 1 2-26.
Massey, D.M., Hayward, AC. and G.W. Winsor. 1 984. Some responses of tomato to
calcium concentration. Ann. Rep. Glasshouse Crops Res. Inst. 1 983 :54-56.
May, D.M. and J. Gonzales. 1 994. Irrigation and nitrogen management as they affect
fruit quality and yield of processing tomatoes. Acta Hort. 376:227-234.
May, D.M. and J. Gonzales. 1 999. Major California processing tomato cultivars
respond differentially in yield and fruit quality to various levels of moisture
stress. Acta Hort. 487 :525-529.
1 6 1
May, D.M. and S . Guillen. 200 1 . Processing varieties that field store to increase grower
and processor profit. Acta Hort. 542:365-372.
May, D.M., D. Peters, T. Wolcott and D.W. Grimes. 1 990. Moisture stress as it affects
yields, soluble solids and viscosity oftomatoes. Acta Hort. 277 : 1 23- 128 .
McCutchan, H. and K.A Shackel. 1 992. Stem-water potential as a sensitive indicator of
water stress in prune trees (Prunus domestica L. cv. French). J . Amer. Soc. Hort.
Sci. 1 1 7 :607-61 1 .
McDonald, A.J.S., T. Ericsson and C.M. Larsson. 1 996. Plant nutrition, dry matter gain
and partitioning at the whole-plant level. J. Exp. Bot. 47: 1 245- 1 253.
McDowell , R.W. and AN. Sharpley. 200 1 . Approximating phosphorus release from
soils to surface runoff and subsurface drainage. 1. Environ. Qual. 30:508-520.
Mencarelli, F. and M.E. Saltveit. 1 988. Ripening of mature-green tomato fruit slices. 1. Amer. Soc. Hort. Sci. 1 1 3 :742-745.
Mengel, K. and E.A Kirkby. 1 987. Principles of plant nutrition. 4th ed. Int. Potash
Inst . , Bern, Switzerland.
Meyer, W.S . and G.c. Green. 1 980. Water use by wheat and plant indicators of average
soil water. Agron. J. 72:253 .
Meyer, W.S. and G.C. Green. 1 98 1 . Plant indicators of wheat and soybean crop water
stress. Irr. Sci. 2 : 1 67- 1 76.
Mitchell, J.P., C. Shennan and S.R. Grattan. 1 99 1 a. Developmental changes in tomato
fruit composition in response to water deficit and salinity. Physiol. Plant.
83 : 1 77- 1 85 .
Mitchell, J.P., C . Shennan, S .R. Grattan and D.M. May. 1 991b. Tomato fruit yields and
quality under water deficit and salinity. J. Amer. Soc. Hort. Sci. 1 1 6:21 5-22 1 .
Murray, M. 1 999. Lower Sacramento valley processing tomato irrigation field test
results. Acta Hort. 487:5 l O-523.
Nichols, M.A, K.J. Fisher and P. 10hnstone. 200 1 . The effect of harvest date on the
yield of processing tomato. Acta Hort. 542:359-363.
Nicklow, C.W. and J.D. Downes. 1 97 1 . Influence of nitrogen, potassium and plant
population on the maturity of field seeded tomatoes for once-over harvest. 1. Amer. Soc. Hort. Sci. 96:46-49.
Olsen, S .R. and L.E. Sommers. 1 982. Phosphorus, pA03-430. In: AL. Page et al. (eds.).
Methods of soil analysis: part 2. chemical and microbiological properties. ASA,
Madison, WI.
1 62
Ozbun, J.L. , CE. Boutonnet, S . Sadik, and P.A. Minges. 1 967. Tomato fruit ripening. 1. Effects of potassium nutrition on occurrence of white tissue. Proc. Amer. Soc.
Hort. Sci . 91 :566-572
Paiva, E.A.S. , H.P. Martinez, V.W.D. Casali and L. Padilha. 1 998. Occurrence of
blossom-end rot in tomato as a function of calcium dose in the nutrient solution
and air relative humidity. J. Plant Nutr. 2 1 :2663-2670.
Pan, H.Y., K.J. Fisher and M.A. Nichols. 1 999. Fruit yield and maturity characteristics
of processing tomatoes in response to drip irrigation. J. Veg. Crop Prod. 5( 1 ) : 1 3-
29.
Phene, C.J. 1 999. Efficient irrigation systems and irrigation scheduling for processing
tomato: the challenge. Acta Hort. 487:479-485.
Phene, C.J., J.M. Miyamoto, K.R Davis, RL. McCormick and RC. Liu. 1 986.
Automated feedback irrigation scheduling and control with a weighing
lysimeter. Proc. Agri-Mation 2 : 1 35-147.
Picha, D.H. and C.B. Hall. 1 98 1 . Influences of potassium, cultivar, and season on
tomato graywall and blotchy ripening. J. Amer. Soc. Hort. Sci. 106 :704-708.
Picken, A.J., K. Stewart and D. Klapwijk. 1 986. Fruit development, p . 1 1 1 - 1 66. In: J.G.
Atherton and J. Rudich (eds.). The tomato crop: a scientific basis for
improvement. Chapman and Hall, New York, N.Y.
Pote, D.H. , T.C Daniel, A.N. Sharpley, P.A. Moore, D.R. Edwards and D.J. Nichols.
1 996. Relating extractable soil phosphorus to phosphorus losses in runoff. Soil
Sci. Soc. Amer. J. 60:855-859.
Pujos, A. and P . Morard. 1 997. Effects of potassium deficiency on tomato growth and
mineral nutrition at the early production stage. Plant and Soil 1 89: 1 89- 1 96.
Pulupol, L.U., M.H. Behboudian and K.J. Fisher. 1 996. Growth, yield and postharvest
attributes of glasshouse tomatoes produced under water deficit. HortScience
3 1 :926-929.
Rahn, C.R 2002. Management strategies to reduce nutrient losses from vegetable crops.
Acta Hort. 57 1 : 1 9-29.
Reid, J.B. and A.R. Renquist. 1 997. Enhanced root production as a feed-forward
response to soil water deficit in field-grown tomatoes. Austral. J. Plant Physiol.
24:685-692.
Rendon-Poblete, E. 1 980. Effect of soil water status on yield, quality and root
development of several tomato genotypes. Univ. of Calif. , Davis, PhD Diss.
1 63
Renquist, AR. and lB. Reid. 1 998. Quality of processing tomato (Lycospersicon
esculentum) fruit from four bloom dates in relation to optimal harvest timing.
N.Z. J. Crop Hort. Sci. 26: 1 6 1 - 1 68.
Renquist, AR. and J.B. Reid. 200 1 . Processing tomato fruit quality: influence of soil
water deficits at flowering and ripening. Austral. J. Agr. Res. 52:793-799.
Rudich, J., D. Kalmar, C. Geizenberg and S. Harel . 1 977. Low water tensions in defined
growth stages of processing tomato plants and their effects on yield and quality.
J. Hort. Sci. 52:391 -399.
Rudich, J., E. Rendon-Poblete, M.A Stevens and A Ambri . 1 98 1 . Use of leaf water
potential to determine water stress in field-grown tomato plants. J. Amer. Soc.
Hort. Sci. 1 06:732-736.
Sakamoto, Y., S . Watanabe, T. Nakashima and K. Okano. 1 999. Effects of salinity at
two ripening stages on the fruit quality of single-truss tomato grown in
hydroponics. J. Hort. Sci. Biotech. 74:690-693 .
Saure, M.C. 200 1 . Blossom-end rot of tomato (Lycospersicon esculentum Mill.) - a
calcium- or a stress-related disorder? Scientia Hortic. 90: 1 93-208.
Schaffer, AA, M. Petreikov, D. Miron, M. Fogelman, M. Spiegelman, Z. Bnei-Moshe,
S. Shen, D. Granot, R. Hadas, N. Dai, H. Levin, M. Bar, M. Friedman, M.
Pilowsky, N. Bilboa and L. Chen. 1 999. Modification of carbohydrate content in
developing tomato fruit. HortScicnce 34: 1 024-1 027.
Sharpley, AN. 1 995. Dependence of runoff phosphorus on extractable soil phosphorus.
J. Environ. Qual. 24:920-926.
Sharpley, AN. 1 997. Rainfall frequency and nitrogen and phosphorus runoff from soil
amended with poultry litter. J. Environ. Qual. 26: 1 1 27 - 1 1 32 .
Sims, J.T. 1 998. Phosphorus soil testing: innovations for water quality protection.
Comm. Soil Sci. Plant Anal. 29: 1 47 1 - 1 489.
Sims, J.T., AC. Edwards, O.F. Schoumans and R.R. Simard. 2000. Integrating soil
phosphorus testing into environmentally based agricultural management
practices. J. Environ. Qual. 29:60-7 1 .
Sinclair, T.R. and T. Horie. 1 989. Leaf nitrogen, photosynthesis, and crop radiation use
efficiency: a review. Crop Sci. 29:90-98.
Stark, J.C., W.M. Jarrell, J. Letey and N. Valoras. 1 983. Nitrogen use efficiency of
trickle-irrigated tomatoes receiving continuous injection of N. Agron. J . 75 :672-
676.
1 64
Stevens, M.A 1 972. Relationships between components contributing to quality
variation among tomato lines. J. Amer. Soc. Hort. Sci. 97:70-73 .
Stevens, M.A 1 976. Inheritance of viscosity potential in tomato. J. Amer. Soc. Hort.
Sci. l 0 l : 1 52- 1 55.
Stevens, M.A and J. Rudich. 1 978. Genetic potential for overcoming physiological
limitations on adaptability, yield and quality in tomato. HortScience 1 3 :673-678.
Strand, L.L. 1998. Integrated pest management for tomatoes. 4th ed. Univ. of Calif. ,
Div. Agr. Natural Resources Publ. 3274, Oakland, CA.
Technicon. 1 973. Potassium, calcium and magnesium are detennined by atomic
emission (K) absorption (Ca and Mg) spectroscopy following digestion in nitric
acid. Industrial method 98/70w. Technicon, New York, N.Y.
Tei, F., P. Benincasa and M. Guiducci. 2002. Critical nitrogen concentration in
processing tomato. Europ. 1. Agron. 1 8:45-55.
Thomas, G.W. 1 982. Exchangeable cations, p . 1 59- 1 65 . In: AL. Page et al. (eds.).
Methods of soil analysis : part 2 . chemical and microbiological properties. ASA,
Madison, WI.
Tregidga G., M. Tregidga and R. Lewthwaite. 1 986. Hydroponics: nutrient film
technique for commercial vegetable growing. AgLink HPP322. Ministry Agr.
Fisheries, Wellington, N.Z.
Trudel, M.J. and J.L. Ozbun. 1 970. Relationship between chlorophylls and carotenoids
of ripening tomato fruit as influenced by potassium nutrition. 1 Exp. Bot.
2 1 :88 1 -886.
Trudel, MJ. and J.L. Ozbun. 197 1 . Influence of potassium on carotenoid content of
tomato fruit. J. Amer. Soc. Hort. Sci. 96 :763-765 .
Twine, 1.R. and C.H. Williams. 197 1 . The determination of phosphorus in Kjeldahl
digests of plant material by automatic analysis. Commun. Soil Sci. Plant Anal.
2 :485-489.
Usherwood, N.R. 1985. The role of potassium in crop quality, p.489-5 1 3 . In: R.S.
Munson (ed.). Potassium in agriculture. ASA-CSSA-SSSA, Madison, WI.
Waisel, Y. 1 996. Aeroponics: a tool for root research, p.239-245. In: Y. Waisel et al.
(eds.). Plant roots: the hidden half. Marcel Dekker, New York, N.Y.
Ward, G.M. 1 964. Greenhouse tomato nutrition - a growth analysis study. Plant Soil
2 1 : 1 25- 1 33.
1 65
Wardlaw, I.F. 1 990. The control of carbon partitioning in plants. New Phytol . 1 1 6:34 1 -
38 l .
Widders, I.E. and O.A. Lorenz. 1 979. Tomato root development as related to potassium
nutrition. J. Amer. Soc. Hort. Sci. 1 04:2 1 6-220.
Widders, I.E. and O.A. Lorenz. 1 982. Potassium nutrition during tomato plant
development. J. Amer. Soc. Hort. Sci. 1 07 :960-964.
Wilcox, G.E. 1 993. Tomato, p. 1 37- 1 4 1 . In: W.E. Bennett (ed.). Nutrient deficiencies
and toxicities in crop plants. AP A, St. Paul, MN. Wolf, S. and J. Rudich. 1 988. The growth rate of fruits on different parts of the tomato
plant and the effect of water stress on dry weight accumulation. Scientia Hort.
34: 1 - 1 1 .
Wudiri, B.B. and D.W. Henderson. 1 985. Effects of water stress on flowering and fruit
set in processing tomatoes. Scientia Hort. 27: 1 89- 1 98 .
Zalom, F.G. and L.T. Wilson. 1 999. Predicting phenological events of California
processing tomatoes. Acta Hort. 487:41 -47
Zegbe-Dominguez, lA., M.H. Behboudian, A. Lang and B.E. Clothier. 2003 . Deficit
irrigation and partial rootzone drying maintain fruit dry mass and enhance fruit
quality in 'Petopride' processing tomato (Lycopersicon esculentum, Mill .) .
Scientia Hort. 98 :505-5 1 0.
166
APPENDICES:
1 67
Appendix I. Summary of low (0.7 dS m·l) and high (2.0 dS m·l) fertility treatments, 1 998-
1999 aeroponic trial.
Nutrient
Nitrogen Phosphorus Potassium
Caclium Magnesium Iron Manganese
Boron Copper Zinc Molybdenum
Treatment regime Z Low fertility High fertility
(mg L-l) (mg L-l)
73 22
1 1 6
59 1 7 4
0 .7
0. 1 1 0.02 0.02 0 .02
208 62
332
1 68 49 1 2
2.0
0 .30 0.07 0.07 0.05
Z high fertility was based on recommendations by Tregidga et al. ( 1 986) for greenhouse tomato production
1 68
.-
() 0 --
Q) L.. ::J
..... ro L.. Q) c.. E Q)
..... L..
<{
35
30
25
20
1 5
1 0
5
0
e
Dec 1 6 Dec 30 Jan 1 3 Jan 27 Feb 1 0 Feb 24 Mar 9 Mar 23 Apr 6 Apr 20
Date
Appendix II-a. Minimum (0) and maximum (e ) daily air temperature during summer trial,
1999-2000. Measurements taken at AgResearch Grasslands Centre, Tennent Road
(less than 1 km from the experimental site).
.-
() 0 --
Q) L.. ::J
..... ro L.. Q) c.. E ID
.....
·0 Cl)
25
20 -
1 5 -
1 0 -
5 -
� ) co lb
�o ;fi (Voi \lJtth1V"� V lit 1 0
01 � J � V rl �(jcr5D"b � � \o�o
O�--�----r---.---�----�---.--�'---'---�----� Dec 16 Dec 30 Jan 1 3 Jan 27 Feb 1 0 Feb 24 Mar 9 Mar 23 Apr 6 Apr 20
Date
Appendix II-b. Mean daily soil temperature at 15 cm during summer trial, 1999-2000.
Measurements taken at AgResearch Grasslands Centre, Tennent Road (less than 1
km from the experimental site).
1 69
1 0�-------------------------------------------------------,
8
- 6 E E -
et ill 4
2
o
. . . . . . . . . . . . . ���?!'�! . . mean
Dec 1 6 Dec 30 Jan 1 3 Jan 27 Feb 1 0 Feb 24 Mar 9 Mar 23 Apr 6 Apr 20
Date
Appendix II-c. Daily pan evaporation (Ep) during summer trial, 1999-2000. Measurements
taken at AgResearch Grasslands Centre, Tennent Road (less than 1 km from the
experimental site).
-E 30 E .........
� c: 20 m 0::
1 0
o
-
-
_Ul I I1 lit. .I 11. ' . I I I I I I Dec 1 6 Dec 30 Jan 1 3 Jan 27 Feb 1 0 Feb 24 Mar 9 Mar 23 Apr 6 Apr 20
Date
Appendix "-d. Daily rainfall during summer trial, 1999-2000. Measurements taken at
AgResearch Grasslands Centre, Tennent Road (less than 1 km from the experimental
site).
1 70
1 2
(a)
1 0
8 ..-
E �e�0!lal _ _
E 6 mean
---
0
� W 4
2
0
1 2
(b)
10
8 ..-
E E 6
�e�0!l"I _ _ _ _
--- mean 0
� W 4
2
0
Apr 24 May 8 May 22 Jun 5 Jun 1 9 Jul 3 Jul 1 7 Jul 3 1 Aug 1 4 Aug 28
Date
Appendix Il l-a. Daily reference evapotranspiration (ETo) during summer trials, 2001 (a)
and 2002 (b). Measurements taken at elMIS Station #6 (less than 1 km from the
experimental sites).
1 7 1
..-() o -..-� ::J ..... � Q) c.. E Q) ..... � <c
..-() o -..-� ::J ..... CO � Q) c.. E Q) ..... � <c
40 _" (a)
" -1 � n • � J � • � � J � � �
30i · , , , t f� r \ I � L M 1.1
V ' O�----r----T----T---�-----r----T---�----�----r---�
40 -" (b)
30
o���APVvl��-t¥�J 20 -
1 0 -i
o i i I I I I I i i i ! Apr 24 May 8 May 22 Jun 5 Jun 1 9 Jul 3 Jul 1 7 Jul 31 Aug 1 4 Aug 28
Date
Appendix Ill-b. M inimum (0) and maximum ( .) daily air temperature during summer trials,
2001 (a) and 2002 (b). elMIS Station #6 (less than 1 km from the experimental s ite).
Gaps indicate missing sensor data.
1 72
-() 0 ---� :J ...... ro L-a> a. E a> ......
0 en
-() 0 ---a> L-:J ...... ro L-a> a. E a> ......
·0 (J)
25
20
1 5
1 0 -
5 -
0
25
20 -
1 5 -
1 0 -
5
(a)
J'� �� � o� J�
/a/'�
I I I I
(b) �£��&"J"""o I csrf d
Ob �b g %, 6' 1t if
O�---T----T-,---r----r---�--�----'----T----�,---�, Apr 24 May 8 May 22 Jun 5 Jun 19 Jul 3 Jul 1 7 Jul 31 Aug 1 4 Aug 28
Date Appendix Ill-c. Mean daily soil temperature at 1 5 cm during summer trials, 2001 (a) and
2002 (b). Measurements taken at elMIS Station #6 (less than 1 km from the
experimental sites). Gaps indicate missing sensor data.
1 73
Appendix IV. Managing fruit soluble solids with late-season deficit irrigation in drip
i rrigated processing tomato production (HortScience 40: 1857-1861 ).
1 74
Top Related